Re: Bruce Schneier's Proposal to dedicate November meeting to saving the Internet from the NSA
On Fri, Sep 6, 2013 at 6:02 PM, Tim Bray tb...@textuality.com wrote: How about a BCP saying conforming implementations of a wide-variety of security-area RFCs MUST be open-source? *ducks* And the user MUST compile them themselves from the sources? Nobody runs open source, (unless its an interpreted language). They run the compiled version and there is no infrastructure to check up on the compilation. And don't forget: http://cm.bell-labs.com/who/ken/trust.html Ned
Re: stability of iana.org URLs
Hi, The link in RFC3315 is actually incorrect -- it should have been http://www.iana.org/assignments/enterprise-numbers, without the file extension, and there's an erratum about this. HTML was generally (if not exclusively) reserved for files that needed to include links to registration forms . As Barry said, we do intend to keep that same short http://www.iana.org/assignments/example format working for every current page, even the newly-created ones. We also prefer to see that format used in documents, since we can't guarantee that the file extension used for the long version won't change. (This information will be appearing on the website in some form.) Also, if you find that a formerly-valid URL (like one that used to have an .html exception) isn't redirecting to the current page, please report it to i...@iana.orgmailto:i...@iana.org. A redirect should have been set up. Excellent, thanks. Ned
Re: stability of iana.org URLs
On 7/31/13 4:06 PM, Barry Leiba wrote: I just followed http://www.iana.org/assignments/enterprise-numbers.html From RFC3315 (DHCPv6)'s reference section. Ten years later, the URL doesn't work. I know that things were reworked when we went to XML based storage, but I thought that the old URLs would at least have a 301 error on them. I discovered that dropping the .html gets me the right data at: http://www.iana.org/assignments/enterprise-numbers Yes: that's the form that IANA would like you to use. They changed their registries from HTML to XML, and the URLs changed. That's true, but cool URIs don't change: http://www.w3.org/Provider/Style/URI.html IMHO it'd be easy enough to put general redirects in place. +1. mod_rewrite is nobody's friend, but it can be tamed and put to good use. Ned
Re: Content-free Last Call comments
Dave Cridland wrote: I strongly feel that positive statements have value, as they allow the community to gauge the level of review and consensus, and I suspect that human nature means that we get more reviews if people get to brag about it. Agreed 100%. But also consider the likely effect of calling certain comments useless. Discussions like this don't exactly fire me up with enthusiasm to expend additional time reviewing and commenting on documents not directly related to what I do. And I rather doubt I'm alone in this. Ned
Re: Language editing
Maybe things have changed, but, if one actually believes the robustness principle, then, in the case Geoff cites, Exchange is simply non-conforming -- not because the spec prohibits rejecting on the basis of a fine distinction about IPv6 formats, but because doing so is unnecessary, inconsistent with the robustness principle, and, arguably, plain silly. I'm afraid I'm going to have to disagree here. If you look at RFC 5321 and are unaware of the history of how the text came about, it gives the definite appearance of going out of its way to ban the use of :: to replace a single 0. A reasonable interpretation is therefore that such forms are disallowed for a reason. It's fine to be tolerant of stuff the relevant standard doesn't allow but doesn't call out explicitly. But that's not the case here. In any case, if you want to fix this, we could change RFC 5321 to accept this form. But as Mark Andrews points out, you can't make it legal to send such forms without breaking interoperability. I suppose we could make the change and recycle at proposed, but that seems rather extreme to fix what is in fact a nonissue. I'll also point out that this has diddley-squat to do with formal verification processes. Again as Mark Anrdrews points out, we deployed something with a restriction that subsequently turned out to be unnecessary, and now we're stuck with it. Indeed, had formal verification processes been properly used, they would have flagged any attempt to change this as breaking interoperability. Ned
Re: Language editing
On 08/05/2013 03:28, John C Klensin wrote: ... I'll also point out that this has diddley-squat to do with formal verification processes. Again as Mark Anrdrews points out, we deployed something with a restriction that subsequently turned out to be unnecessary, and now we're stuck with it. Indeed, had formal verification processes been properly used, they would have flagged any attempt to change this as breaking interoperability. Also agreed. To be clear, I'm no fan of formal verification either, but this *is* a case where the IETF's lapse in formality has come back to bite, and the Postel principle would have helped. Also, given the original subject of the thread, I don't see how language editing could have made any difference. Reread the notes about the history behind this in this thread. You haven't even come close to making a case that formal verification of the standards would have prevented this from happening. (Formal verification of implementation compliance to the standards would of course have prevented Apple's client bug, but that's a very different thing.) You are, however, correct that this has nothing to do with specification editing. Ned
Re: Language editing
On 08/05/2013 08:33, Ned Freed wrote: On 08/05/2013 03:28, John C Klensin wrote: ... I'll also point out that this has diddley-squat to do with formal verification processes. Again as Mark Anrdrews points out, we deployed something with a restriction that subsequently turned out to be unnecessary, and now we're stuck with it. Indeed, had formal verification processes been properly used, they would have flagged any attempt to change this as breaking interoperability. Also agreed. To be clear, I'm no fan of formal verification either, but this *is* a case where the IETF's lapse in formality has come back to bite, and the Postel principle would have helped. Also, given the original subject of the thread, I don't see how language editing could have made any difference. Reread the notes about the history behind this in this thread. You haven't even come close to making a case that formal verification of the standards would have prevented this from happening. You are correct if only considering the mail standards. I suspect that a serious attempt at formal verification would have thrown up an inconsistency between the set of mail-related standards and the URI standard. Which is relevant to the present situation... how exactly? And in any case, the relevant URI standard incorporates the ABNF from RFC 2373, but doesn't state whether or not it also inherits restrictions specified in prose in that specification, which is where the restriction in RFC 2821 originated. However, I think the underlying problem here is that we ended up defining the text representation of IPv6 addresses in three different places, rather than having a single normative source. (ABNF in the mail standards, ABNF in the URI standard, and English in ipv6/6man standards.) Except that wasn't the problem. The ABNF in the email standards is consistent with what the other standards said when RFC 2821 was published. And once that happened the die was cast as far as email usage was concerned. The fact that the other standards later decided to loosen the rules in this regard is what caused the inconsistency. If you want to blame something, it has to be either the initial decision to limit use of :: or the subsequent decision to remove that limit. And for increased formalism to have helped it would have to have prevented one of those from happening. I supose that's possible, but I certainly don't see it as inevitable. Ned
Re: Language editing
Mark Andrews ma...@isc.org wrote: Apples mail client is broken [IPv6:2001:df9::4015:1430:8367:2073:5d0] is not legal according to both RFC 5321 and RFC 2821 which is all that applies here. I was until today unaware how strong the feelings are on this one-or-more vs. two-or-more issue. I do not expect to change anybody's mind. :^( But I do object to calling that EHLO string not legal. It's syntactically illegal according to the ABNF in RFC 5321 itself, which specifically states: IPv6-comp = [IPv6-hex *5(: IPv6-hex)] :: [IPv6-hex *5(: IPv6-hex)] ; The :: represents at least 2 16-bit groups of ; zeros. No more than 6 groups in addition to the ; :: may be present. The 5321 reference names RFC 4291 as the source of address syntax (even if it gives BNF which says two or more if you delve deeply enough). It may be the source, but the formal syntax specified in the document at hand has to be the definitive one. RFC 4921 is clear about saying one or more. The Errata posted against it claiming it should say two or more have been rejected. It is silly to argue under these conditions that Apple's EHLO string is not legal. No, what's silly is your argument that the ABNF in the actual specification of how mail clients and servers are supposed to behave isn't the definitive definition. BTW, RFC 5321 still contains the language about if the verification fails, the server MUST NOT refuse to accept a message on that basis. And that language appears in the context of checking that the IP literal in the HELO/EHLO actually corresponds to the IP address of the client. It has nothing to do with syntactic validity. so IMHO enforcing any particular interpretation of what an IPv6 address literal should look like is double-plus-ungood. Then you should be arguing for a change in RFC 5321, because it is *very* clear that this usage is not allowed. Ned
Re: Internet Draft Final Submission Cut-Off Today
On Feb 26, 2013, at 5:38 PM, Pete Resnick presn...@qti.qualcomm.com wrote: But more seriously: I agree with you both. The deadline is silly. +1 +1 The deadline originated because the secretariat needed time to post all of those drafts (by hand) before the meeting. The notion of an automated tool that blocks submissions for two weeks before the meeting is just silly. All the more so since it just leads people to use informal distribution methods. I don't recall a case where a chair forbid the discussion of a draft distributed this way. I recall hearing something once about routing around obstacles... Pity we don't internalize such principles fully. Ned
Re: Internet Draft Final Submission Cut-Off Today
On 02/27/2013 01:49 PM, Carsten Bormann wrote: On Feb 27, 2013, at 19:18, ned+i...@mauve.mrochek.com wrote: routing around obstacles It turns out for most people the easiest route around is submitting in time. That is actually what counts here: how does the rule influence the behavior of people. Chair hat: WORKSFORME. (And, if I could decide it, WONTFIX.) +1. As far as I can tell, the deadline actually serves the purpose of getting people to focus on IETF and update their documents sufficiently prior to the meeting, that it's reasonable to expect meeting participants to read the drafts that they intend to discuss. And I say this as someone who, as an author, has often found the deadline to be very inconvenient. And your evidence for this is .. what exactly? Yes, the deadline makes the drafts show up a bit sooner, but I rather suspect that the overwhelming majority of people don't bother to do much reading in the inverval. I certainly don't. And given the ready available tools to tell the reader what's changed I don't need to. In almost all cases for -nn where nn 00 I can check what's changed in a few minutes, and I can do it in the context of the actual work being done. I don't really have any objection to a -00 cutoff, but the second cutoff is nothing short of asinine. In any case, I personally have basically stopped caring about the deadline and I encourage others to do the same. If I make the deadline fine, if not I post the update somewhere else, done. Ned
Re: The RFC Acknowledgement
I try to include in the Acknowledgements section of any Internet Drafts I edit the names of anyone who comments on the draft if (1) the comment results in a change in the draft and (2) the commenter does not request that they be left out. If you comment on some draft and the draft is changed as a result and you want to be acknowledged and you are not added to the acknowledgements list, you should complain to the editor / author. That's exactly the policy Nathaniel Borenstein and I agreed to use for MIME. I've used it ever since for all the documents I have edited, and it seems to have worked well. (And apologies to anyone whose name I have omitted under that policy - if I did that it was entirely inadvertent.) The only time I've ever had an acknowledgments section has been when an author or contributor is deceased. This very unfortunate situation is quite delicate and merits handling on a case-by-case basis; IMO no specific policy could possibly be written to accomodate it. Ned
Re: Vestigial Features (was Re: CRLF (was: Re: A modest proposal))
On Jan 24, 2013, at 04:41, wor...@ariadne.com (Dale R. Worley) wrote: From: Carsten Bormann c...@tzi.org I think in protocol evolution (as well as computer system evolution in general) we are missing triggers to get rid of vestigial features. That's quite true. Let us start by rationalizing the spelling and punctuation of written English (which is the coding system for *this entire discussion*). Once we've cleaned up that idiocy, we can start in on SIP. I see I didn't make myself clear. I'm not suggesting we clean up vestigials in existing spec[ie]s, such as HTTP or SIP. (We might even do that, see HTTPbis, but only very carefully.) My point was about the case when we clone new stuff off existing protocols. (SIP was cloned off HTTP which was cloned off SMTP which was cloned off FTP which at least has had strong kinship with Telnet, hence all these use NVT.) Actually, in regards to NVT, there's something of a break in HTTP: text material is *not* required to use CRLF as a line terminator. Any of LF, CR, or CRLF is permissible. I really don't want to get into the history of this choice or for that matter the results it has produced. Suffice it to say it has had some advantages and some disadvantages. Staying in your analogy: when you design a new language based on English, please do fix some of this stuff. (Maybe the analogy isn't that useful after all.) Actually, we haven't been that bad about this. E.g., we have been pretty good about getting rid of the madness of historic character coding by focusing on UTF-8 for new designs. But a vital part of that choice is that it is backwards compatible with US-ASCII. What I'm asking for: Apart from these big reformation projects, we also should occasionally fix little things. Or not. How about instead of viewing these sorts past choices as things to be fixed, we instead view them as what they were: Choices. And then decide, based on the actual requirements of whatever we're doing, whether or not it makes sense to make a change. Sure, we can design some new format with LFs or CRs or maybe even line or page separators as line breaks instead of CRLFs. And maybe that makes sense in some cases. Or not: Such choices can have serious consequences if any degree of backwards compatibility with exiting transfer services is desired. And constant transcoding of stuff to make it work in various places is not necessarily a winning approach. When inventing new stuff by drawing analogies off of old specs. I really don't think unthinking copying from past specifications is the issue here, your asssertions to the contrary notwithstanding. Ned
Re: Last Call: draft-farrell-ft-03.txt (A Fast-Track way to RFC with Running Code) to Experimental RFC
Hi Ned, On 01/16/2013 03:40 AM, Ned Freed wrote: Actually I think you make a couple of great points that ought be mentioned in the draft about implementability. (No chance you'd have time to craft a paragraph? If not, I'll try pinch text from above:-) Now that you point it out like that, I'm irritated at myself for not having included it already! (It rings bells for me.) OK, I think the place for a new paragraph is just before the last paragraph of section 2. How about something along the lines of: A complete and correct specification is not in and of itself a guarantee of high quality implementations. What may seem like minor details can increase implementation difficulty substantially, leading to implementations that are fragile, contain unnecessary restrictions, or do not scale well. Implementation experience has the potential to catch these problems before the specification is finalized and becomes difficult to change. You might also want to change the final paragraph of the section a bit in light of the addition; I'll leave that to you to work out. Did that, working copy at [1]. Lemme know if there're any changes that are needed. It looks good to me. I also note that there's a reference to interoperability in the fourth paragraph of section 1. Perhaps changing For example, a framework draft will not be a good candidate because implementations of such documents are not, of themselves, interoperable. to something like For example, a framework draft will not be a good candidate because implementations of such documents are incomplete and therefore do not demonstrate either implementability or interoperability of an entire protocol. would be in order. Ned
Re: Last Call: draft-farrell-ft-03.txt (A Fast-Track way to RFC with Running Code) to Experimental RFC
Martin Rex wrote: John Leslie wrote: I'm pretty darn uncomfortable _ever_ picking a fight with any sitting AD, But I feel obligated to say this seems like a terrible idea to me. As a background, I'm a long-time believer in rough consensus for Proposed Standard and running code for advancement along the standards track. I do not believe the two mix well. I don't have the resource to participate the discussion, but these statements capture my opinion pretty well. I'm reluctantly going to have to join John and Martin, with one very important difference: I think there is huge value in implementing specifications during the process leading up to proposed standard. (As many people know, I do this myself quite often for drafts I'm interested in.) I would rather phrase it as having demonstrable interoperability for advancement, and ideally having implementations done much earlier. More specifically, where I part ways with this draft is in believing that very early implementation helps find interoperability issues. I have not found that to be the case. What it helps find are issues affecting implementability. In the applications area at least, it's surprisingly easy to specify things that look good on paper but turn out to suck when you try to code them. (My own sins in this area are well known - RFC 2231 - and one of the reasons for that it is one of the few RFCs I've written without implementing it first.) Now, it's quite true that implementability problems can lead to interoperability problems, like when different people take different shortcuts in coding an unnecessarily problematic specification. But other outcomes are possible, including fragility, unnecessary restrictions, and most especially scalability problems. Another problem with focusing on interoperability specifically as opposed to overall implementability is that it makes it hard to argue that one or two implementations provide that much benefit. In my experience having just one implementation, even one done by a draft author, is surprisingly helpful in cleaning up drafts. I guess what I would like to see is for the draft to talk a little more about finding implementation issues in general and a lot less about finding interoperability issues specifically. I also think the draft goes a bit far in the carrots it provides, but that may have more to do with my own experiences in the applications area, where important comments have a way of only showing up at the last second and where therefore the abbreviated process might be a little dangerous to use. Finally, I'm going to apologize for the tardiness of this comment, which really should have been made sooner. I'm also going to apologize in advance for probably not being able to fully participate in this discussion due to severe time constraints. Ned
Re: Last Call: draft-farrell-ft-03.txt (A Fast-Track way to RFC with Running Code) to Experimental RFC
Actually I think you make a couple of great points that ought be mentioned in the draft about implementability. (No chance you'd have time to craft a paragraph? If not, I'll try pinch text from above:-) Now that you point it out like that, I'm irritated at myself for not having included it already! (It rings bells for me.) OK, I think the place for a new paragraph is just before the last paragraph of section 2. How about something along the lines of: A complete and correct specification is not in and of itself a guarantee of high quality implementations. What may seem like minor details can increase implementation difficulty substantially, leading to implementations that are fragile, contain unnecessary restrictions, or do not scale well. Implementation experience has the potential to catch these problems before the specification is finalized and becomes difficult to change. You might also want to change the final paragraph of the section a bit in light of the addition; I'll leave that to you to work out. Hope this helps. Ned
Re: A proposal for a scientific approach to this question [was Re: I'm struggling with 2219 language again]
On 07/01/2013 12:42, Stewart Bryant wrote: Indeed an interesting additional question. My view is that you MUST NOT use RFC2119 language, unless you MUST use it, for exactly that reason. What is important is on the wire (a term that from experience is very difficult to define) inter-operation, and implementers need to be free to achieve that though any means that suits them. Agreed. Imagine the effect if the TCP standard had said that a particular congestion control algorithm was mandatory. Oh, wait... ... RFC 1122 section 4.2.2.15 says that a TCP MUST implement reference [TCP:7] which is Van's SIGCOMM'88 paper. So apparently any TCP that uses a more recent congestion control algorithm is non-conformant. Oh, wait... ... RFC 2001 is a proposed standard defining congestion control algorithms, but it doesn't update RFC 1122, and it uses lower-case. Oh, wait... RFC 2001 is obsoleted by RFC 2581 which obsoleted by RFC 5681. These both use RFC 2119 keywords, but they still don't update RFC 1122. This is such a rat's nest that it has a guidebook (RFC 5783, Congestion Control in the RFC Series) and of course it's still an open research topic. Attempting to validate TCP implementations on the basis of conformance with RFC 2119 keywords would be, well, missing the point. I know this is an extreme case, but I believe it shows the futility of trying to be either legalistic or mathematical in this area. Exactly. Looking for cases where the use/non-use of capitalized terms caused an interoperability failure is a bit silly, because the use/non-use of such terms doesn't carry that sort of weight. What does happen is that implementation and therefore interoperability quality can suffer when standards emphasize the wrong points of compliance. Things work, but not as well as they should or could. A fairly common case of this in application protocols is an emphasis on low-level limits and restrictions while ignoring higher-level requirements. For example, our email standards talk a fair bit about so-called minimim maximums that in practice are rarely an issue, all the while failing to specify a mandatory minimum set of semantics all agents must support. This has led to lack of interoperable functionality in the long term. Capitalized terms are both a blessing and a curse in this regard. They make it easy to point out the really important stuff. But in doing so, they also make it easy to put the emphasis in the wrong places. tl;dr: Capitulized terms are a tool, and like any tool they can be misused. Ned
Re: I'm struggling with 2219 language again
Dean, I am struggling constantly with 2119 as an AD, because if I take the letter (and the spirit) of 2119 at face value, a lot of people are doing this wrong. And 2119 is a BCP; it's one of our process documents. So I'd like this to be cleared up as much as you. I think there is active harm in the misuse we are seeing. To Ned's points: On 1/4/13 7:05 PM, ned+i...@mauve.mrochek.com wrote: +1 to Brian and others saying upper case should be used sparingly, and only where it really matters. If even then. That's the entire point: The terms provide additional information as to what the authors consider the important points of compliance to be. We will like end up in violent agreement, but I think the above statement is incorrect. Nowhere in 2119 will you find the words conform or conformance or comply or compliance, and I think there's a reason for that: We long ago found that we did not really care about conformance or compliance in the IETF. What we cared about was interoperability of independently developed implementations, because independently developing implementations that interoperate with other folks is what makes the Internet robust. Importantly, we specifically did not want to dictate how you write your code or tell you specific algorithms to follow; that makes for everyone implementing the same brittle code. Meh. I know the IETF has a thing about these terms, and insofar as they can lead to the use of and/or overreliance on compliance testing rather than interoperability testing, I agree with that sentiment. OTOH, when it comes to actually, you know, writing code, this entire attitude is IMNSHO more than a little precious. Maybe I've missed them, but in my experience our avoidance of these terms has not resulted in the magical creation of a widely available perfect reference implementation that allows me to check interoperability. In fact in a lot of cases when I write code I have absolutely nothing to test against - and this is often true even when I'm implementing a standard that's been around for many years. In such cases the use of compliance language - and yes, it is compliance language, the avoidance of that term in RFC 2119 notwithstanding - is essential. And for that matter it's still compliance language even if RFC 2119 terms are not used. I'll also note that RFC 1123 most certainly does use the term compliant in regards to capitalized terms it defines, and if nitpicking on this point becames an issue I have zero problem replacing references to RFC 2119 with references to RFC 1123 in the future. All that said, I'll again point out that these terms are a double-edged sword, and can be used to put the emphasis in the wrong place or even to specify downright silly requirements. But that's a argument for better review of our specifications, because saying MUST do this stupid and couterproductive thing isn't fixed in any real sense by removing the capitalization. The useful function of 2119 is that it allows us to document the important *behavioral* requirements that I have to be aware of when I am implementing (e.g., even though it's not obvious, my implementation MUST send such-and-so or the other side is going to crash and burn; e.g., even though it's not obvious, the other side MAY send this-and-that, and therefore my implementation needs to be able to handle it). And those even though it's not obvious statements are important. It wastes my time as an implementer to try to figure out what interoperability requirement is meant by, You MUST implement a variable to keep track of such-and-so-state (and yes, we see these in specs lately), and it makes for everyone potentially implementing the same broken code. Good point. Pointing out the nonobvious bits where things have to be done in a certain way is probably the most important use-case for these terms. Ned
Re: I'm struggling with 2219 language again
+1 to Brian and others saying upper case should be used sparingly, and only where it really matters. If even then. That's the entire point: The terms provide additional information as to what the authors consider the important points of compliance to be. The notion (that some have) that MUST means you have to do something to be compliant and that a must (lower case) is optional is just nuts. In some ways I find the use of SHOULD and SHOULD NOT be to be more useful than MUST and MUST NOT. MUST and MUST NOT are usually obvious. SHOULD and SHOULD NOT are things on the boundary, and how boundary cases are handled is often what separated a good implementation from a mediocre or even poor one. If the ARP spec were to say, upon receipt of an ARP request, the recipient sends back an ARP response, does the lack of a MUST there mean the response is optional? Surely not. And if we make it only a SHOULD (e.g., to allow rate limiting of responses - a very reasonable thing to do), does lack of MUST now make the feature optional from a compliance/interoperability perspective? The idea that upper case language can be used to identify all the required parts of a specificition from a compliance/conformance/interoperability perspective is just wrong. This has never been the case (and would be exceeding painful to do), though (again) some people seem to think this would be useful and thus like lots of upper case language. At most it provides the basis for a compliance checklist. But such checklists never cover all the points involved in compliance. Heck, most specifications in toto don't do that. Some amount of common sense is always required. Where you want to use MUST is where an implementation might be tempted to take a short cut -- to the detriment of the Internet -- but could do so without actually breaking interoperability. A good example is with retransmissions and exponential backoff. You can implement those incorrectly (or not at all), and still get interoperability. I.e., two machines can talk to each other. Maybe you don't get good intereoperability and maybe not great performance under some conditions, but you can still build an interoperabile implementation. IMO, too many specs seriously overuse/misuse 2119 language, to the detriment of readability, common sense, and reserving the terms to bring attention to those cases where it really is important to highlight an important point that may not be obvious to a casual reader/implementor. Sadly true. Ned
Re: WCIT outcome?
From: John Day jeanj...@comcast.net I remember when a modem came with an 'acoustic coupler' because connecting it directly to the phone line was illegal. No, there was nothing illegal about it. The reason for acoustic couplers was that the RJ-11 had been invented yet and it was a pain to unscrew the box on the wall and re-wire every time you wanted to connect. ... It may have been illegal in some countries but certainly not in the US. Huh? Remember the Carterphone decision? Absolutely. Too bad the FCC didn't see fit to extend it to wireless. The one that overturned FCC Tariff Number 132: No equipment, apparatus, circuit or device not furnished by the telephone company shall be attached to or connected with the facilities furnished by the telephone company, whether physically, by induction or otherwise. Now, your point about rewiring the jack may in fact be the reason for _post-Carterphone_ acoustic couplers, but it was indeed at one time illegal to connect directly (other than AT+T/WE supplied equipment). I'm skeptical about this last part. Prior to the advent of RJ-11 Bell System line cords used a large polarized four pin jack. After Carterphone all sorts of stuff started to appear to accomodate these, including extension cords, plug-jack passthroughs, and even cube taps. At one point there was something that said one phone in each home had to be directly wired without a plug. I don't know if this was a regulation, a phone company rule, or just a suggestion, but it also fell by the wayside after Carterphone. I certainly saw acoustic coupled equipment in use long after Carterphone, but in my experience it was because of general intertia/unwillingness to do the necessary engineering, not because of the lack of connectors. Ned
Acoustic couplers (was: Re: WCIT outcome?)
On 1/2/2013 1:34 PM, ned+i...@mauve.mrochek.com wrote: Now, your point about rewiring the jack may in fact be the reason for _post-Carterphone_ acoustic couplers, but it was indeed at one time illegal to connect directly (other than AT+T/WE supplied equipment). I'm skeptical about this last part. Prior to the advent of RJ-11 Bell System line cords used a large polarized four pin jack. After Carterphone all sorts of stuff started to appear to accomodate these, including extension cords, plug-jack passthroughs, and even cube taps. Acoustic couplers date back farther than the 4-pin plugs. Of course. However, we're talking about post-Carterphone here. Carterphone was 1968, and I'm sure four pin plugs were in use by then. Also keep in mind that ATT fought the Carterphone decision for many years. They got some state regulators to issue their own restrictions, but the FCC nixed them all. Then they said a special protection device had to be used. The FCC shot that down too. They also tried fees, but for that to work people had to tell ATT to charge them, which of course didn't happen. ... At one point there was something that said one phone in each home had to be directly wired without a plug. I don't know if this was a regulation, a phone company rule, or just a suggestion, but it also fell by the wayside after Carterphone. It was usually enforced rigorously. A given field tech might choose to overlook a local mod, but they were authorized to remove such things. So in my apartment, I installed a shutoff switch to the line, to be able to sleep through attempts by my boss to call me in to work an additional shift as a computer operator, at UCLA, around 1970 -- if I answered, I was required to come in. Remember there was no caller ID in those days. The tech who needed to work on my phone service was very clear that he was supposed to remove it. After checking that I had handled the wiring acceptably, he looked at me and said so if I remove this, you'll probably just reinstall it, right? He then left it in place. A line mod was probably against the rules irrespective of Carterphone in those days. But had you bought your own phone with a ringer switch and hooked that up, that absolutely would have been covered by Carterphone. Of course you would then have had to convince ATT of that - see above. Ned
Re: IETF work is done on the mailing lists
So here's my question: Does the community want us to push back on those situations? Does the community believe that the real IETF work is done on the mailing lists, and not in the face-to-face meetings, to the extent that the community would want the IESG to refuse to publish documents whose process went as I've described above, on the basis that IETF process was not properly followed? The issue isn't the lack of comments but any potential lack of opportunity to comment. If the document was announced on the list, prefably including ancillary about changes that have been made, and people chose not to comment there, then that's fine. But if information about the document wasn't made available - as is sometimes the case if the document isn't named under the WG - then that's a problem. Ned
Re: Gen-ART LC review of draft-leiba-5322upd-from-group-06
Minor issues: 1.It is not clear from the draft what the use case for using the group construct is. Section 3 talks about the issues with using the group construct and recommend limited use, but this is the only information. The main driver for this work is to add support for EAI downgrade mechanisms, although a good case can also be made that the present restrictions on group usage are silly, limiting and unnecessary. (The silliness arises from the observation that these restrictions enforced by most agents in the field.) It is far from clear that it makes sense to clutter up this very clean and crisp specification with a bunch of EAI use case scenarios, especially since this is an update to RFC 5322 that could in theory be merged into the main document in the future. A decision taken long ago for both RFC 5321 and 5322 was to keep them as free of entanglements with technology that's layered on top of them (e.g., MIME, various SMTP extensions) as possible. In any case, the decision of the group was not to do this. 2.Section 2.1 says If the sender field uses group syntax, the group MUST NOT contains more than one mailbox. Why use a group name for a single mailbox? Why not? I'm sorry, but this question makes no sense. A group is a characteristic attached to a collection of zero or more addresses. A group containing a single member is every bit as valid as a construct as a group containing 20 members. Or zero, for that matter. Ned
Re: Gen-ART LC review of draft-leiba-5322upd-from-group-06
On 10/17/2012 10:49 AM, ned+i...@mauve.mrochek.com wrote: Minor issues: 1. It is not clear from the draft what the use case for using the group construct is. Section 3 talks about the issues with using the group construct and recommend limited use, but this is the only information. The main driver for this work is to add support for EAI downgrade mechanisms, Although the Intro text cites this, it doesn't explain how the change will help. Nor is this explained elsewhere in the document. A single sentence summarizing what benefit is achieved with the change, along with a couple of usage examples, would go a long way towards showing how this update helps in practical ways. I could live with a single sentence, but I strongly object to the inclusion of examples, for the reasons I gave in my original response. Ned
Re: Gen-ART LC review of draft-leiba-5322upd-from-group-06
--On Wednesday, October 17, 2012 12:00 -0700 ned+i...@mauve.mrochek.com wrote: A single sentence summarizing what benefit is achieved with the change, along with a couple of usage examples, would go a long way towards showing how this update helps in practical ways. I could live with a single sentence, but I strongly object to the inclusion of examples, for the reasons I gave in my original response. Would a possible middle ground be to include a single well-crafted sentence with an informative citation of draft-ietf-eai-popimap-downgrade? That document does contain examples and an explanation of that particular use case. I'd prefer to avoid that entirely for the same reasons Ned cites, but it would, IMO, be good to get on with this rather than quibbling for days or weeks. Channeling my inner Maslow, I see the present text as best, an additional sentence or two as next best, a sentence and a cite to the downgrade doc next in line, and including actual EAI examples in this doc as the worst choice. FWIW. Ned
Re: Gen-ART LC review of draft-leiba-5322upd-from-group-06
On 10/17/2012 2:32 PM, Ned Freed wrote: Channeling my inner Maslow, I see the present text as best, an additional sentence or two as next best, a sentence and a cite to the downgrade doc next in line, and including actual EAI examples in this doc as the worst choice. The problem I have with the current text is that it says 'what' motivated the change, but not how it is useful for the intended class of uses. The reader is left entirely to guess. Self-actualization among the inadequately-informed invites fantasy more than it invites utility. That depends on both the utility and the desirability of repeating the existing use case, doesn't it? The EAI use case of downgraded messages is one that only exists because of a nasty confluence of restrictions and which we absolutely do not want to see repeated in any other context. If you really think this is important to explain why we're making this change against the overall context of RFC 5322 - and I most certainly do not agree that it is important to do so - then the best use case to add is the negative one: The elimination of an unnecessary restriction that isn't followed in practice. I see no way to explain the narrow EAI use case in this context without either dragging in a whole bunch of EAI that has no business being here or leaving various things dangling. Either way the average reader looking at general message usage isn't going to get it, and the more you try and waltz around this the more likely you are to get exactly the outcome you fear: Someone is going to see some sort of parallel with some problem they are trying to solve (some sort of deliberate form of address obfuscation scheme immediately comes to mind) and proceeds to make a mess of things. Ned P.S. It really would be best if this could wait until there was an update to RFC 5322. But that's not how the timing has worked out, so this is the best alternative available.
Re: Gen-ART LC review of draft-leiba-5322upd-from-group-06
Channeling my inner Maslow, I see the present text as best, an additional sentence or two as next best, a sentence and a cite to the downgrade doc next in line, and including actual EAI examples in this doc as the worst choice. The problem I have with the current text is that it says 'what' motivated the change, but not how it is useful for the intended class of uses. The reader is left entirely to guess. So, is it better to put in a sentence about representing non-ASCII text in the group name without including a replyable address? Or is it better to remove the notation about the EAI use case, and just say that it's stupid to have the restriction, so we're removing it? If the alternative is to dig into EAI in any depth at all, the latter is far preferable. Ned
Re: Obsoletes/Updates in the abstract (Was: Gen-ART LC Review of draft-ietf-eai-rfc5721bis-07)
On Sep 21, 2012, at 11:14 AM, Pete Resnick presn...@qualcomm.com wrote: [Changing the subject and removing GenArt and the document authors/chairs] On 9/21/12 10:52 AM, Glen Zorn wrote: -- The abstract should mention that this obsoletes 5721 Why? There is a statement in the header, 10 lines above the abstract, that says Obsoletes: 5721 (if approved). The IESG put this into the nits check before my time. The Last Call and publication announcements normally contain only the abstract, not the metadata above, and I believe the thinking was that if you are a person who scans through those announcements, you probably would (and would want to) take notice of documents that purport to obsolete or update document that you recognize. We could probably change the tool to add the metadata to the announcements, but apparently quite a few people read abstracting services that grab the abstracts of newly published documents. Not much we can do for them. It's certainly useful to some folks. Necessary? (*Shrug*) Not enough wasted bits for me to care one way or the other. As a Gen-ART reviewer, I called it out for exactly the reasons Pete mentions, and care about the same amount :-) But putting it there seems to hurt nothing, and maybe help just a little bit in some cases. That's your opinion. Others, myself included, strongly disagree. Ned
Re: Gen-ART LC Review of draft-ietf-eai-simpledowngrade-07
On 09/19/2012 04:24 AM, Ben Campbell wrote: I am the assigned Gen-ART reviewer for this draft. For background on Gen-ART, please see the FAQ at http://wiki.tools.ietf.org/area/gen/trac/wiki/GenArtfaq . Please resolve these comments along with any other Last Call comments you may receive. Document: draft-ietf-eai-simpledowngrade-07 Reviewer: Ben Campbell Review Date: 2012-09-18 IETF LC End Date: 2012-09-20 Summary: This draft is mostly on the right track, but has open issues Major issues: -- I'm concerned about the security considerations related to having a mail drop modify a potentially signed message. ... Hm, sounds like a misunderstanding. Did you understand that the modification happens in RAM, and that the message stored unmodified and has the valid signature? If not I suppose extra verbiage is needed. I think this is already pretty clear: The texts says things like present and serve. It takes a pretty deliberate misreading to miss those words. Moreover, stating that this happens only in RAM may not be correct. A server could choose to store a downgraded version so it doesn't have to be rerendered, for example. The point is that version isn't going to be presented to a client that supports EAI, not that it only happens in RAM. The signature issue has been discussed. The answer is more or less: The WG expects EAI users to use EAI-capable software, and to accept partial failure when using software that cannot be updated. This entire draft is draft is about damage limitation when an EAI user uses EAI-ignorant software (e.g. your phone, if you do your main mail handling on a computer but occasionally look using the phone). That there will be damage is expected and accepted. IMO it's unavoidable. I think that's more a matter of fact, not opinion. If your software is incapable of presenting something, it's incapable of presenting something. The only question is whether you get a downgraded version the software is capable of handling or nothing. EAI allows that choice to be made by the implementation or operationally. Minor Issues: -- It's not clear to me why this is standards track rather than informational. I don't remember. Perhaps because it needs to update 3501. -- section 3, 2nd paragraph: Are there any limits on how much the size can differ from the actual delivered message? Can it be larger? Smaller? It's worth commenting on whether this could cause errors in the client. (e.g. Improper memory allocation) An input message can be constructed to make the difference arbitrarily large. For instance, just add an attachment with a suggested filename of a million unicode snowmen, and the surrogate message will be several megabyte smaller than the original. Or if you know that the target server uses a long surrogate address format, add a million short Cc addresses and the surrogate will be blown up by a million long CC addresses. I doubt that this is exploitable. You can confuse or irritate the user by making the client say downloading 1.2MB when the size before download was reported as 42kb, that's all. I wish all my problems were as small. If it's exploitable it almost certainly is what I refer to as the tiny twig on the attack tree. That is, if there's a size-related issue there are going to be much easier ways to exploit it than this. I suppose some folks may believe that describing these sorts of twigs is valuable. I disgree and believe it to be unncesssary clutter that is likely to end up detracting from overall security. I'll add a comment and a reminder that the actual size is supplied along with the literal during download. -- Open Issues section: Should Kazunori Fujiwara’s downgrade document also mention DOWNGRADED? Good question. It seems like they should be consistent on things like this. (This is really more a comment on that draft than this one.) I think I've made up my mind that in this case it doesn't matter. Kazunori's task is complex reversible downgrade and has the Downgraded-* header fields, why then bother with the DOWNGRADED response code? But it's not my decision. I agree it doesn't matter. -- Abstract should mention that this updates 3501 Really? A detail of this document updates a minor detail of that document, that's hardly what I would expect to see in a single-paragraph summary. Exactly right. This is a silly thing to do. I know someone who likes to repeat the Subject in the first line of the email body text. Just in case I didn't see it the first time, I suppose. The policy for EAI documents, which was agreed to by the working group and which is now reflected in a number of RFCs, is not to put this sort of nonsense in the abstract. Ned
Re: Draft IESG Statement on Removal of an Internet-Draft from the IETF Web Site
On 9/9/12 8:43 PM, John Levine wrote: Let's say I write to the IESG and say this: Due to a late night editing error, draft-foo-bar-42 which I submitted yesterday contains several paragraphs of company confidential information which you can easily see are irrelevant to the draft. My boss wants it taken down pronto, even though he realizes that third parties may have made copies of it in the meantime. I will probably lose my job if it stays up for more than a few days. Thanks for your consideration. Is this the response? You didn't make any legal threats, and now that we know the situation, we wouldn't believe any legal threats you might make in the future, so you better check out those burger flipping opportunities. No, the response is that we refer you to our policy. As an open organization we do not remove information once posted, except under extraordinary circumstances. Exactly. This sort of thing is wh a policy is needed, although I note in passing that the folks at this hypothetical might want to read up on the Streisand Effect. What was wrong with the original version which gave the IESG the latitude to remove an I-D if they feel, for whatever reason, that it would be a good idea to do so? What original? The draft policy states: An I-D will only be removed from the public I-D archive in compliance with a duly authorized court order. If the IESG were so screwed up that they started deleting I-Ds for bad reasons, no amount of process verbiage would help. Certainly, but let's not start from the wrong place to begin with. Let's also set expectations that the IESG may be used to clean up after other peoples' messes. They have enough to do. That is if anything an understatement. And again, this is best developed with counsel. A very emphatic +1 to this. Ned
Re: Draft IESG Statement on Removal of an Internet-Draft from the IETF Web Site
Let's say I write to the IESG and say this: Due to a late night editing error, draft-foo-bar-42 which I submitted yesterday contains several paragraphs of company confidential information which you can easily see are irrelevant to the draft. My boss wants it taken down pronto, even though he realizes that third parties may have made copies of it in the meantime. I will probably lose my job if it stays up for more than a few days. Thanks for your consideration. Exactly. This sort of thing is wh a policy is needed, although I note in passing that the folks at this hypothetical might want to read up on the Streisand Effect. Note that I phrased it as a polite request, not a threat. I don't see that as especially relevant: There have been plenty of cases where a polite request called attention to something that would otherwise have been ignored, although of course the ones that get reported at http://www.thestreisandeffect.com/ tend to be the ones where bad behavior was involved. And again, this is best developed with counsel. A very emphatic +1 to this. Sure, but keep in mind that it's one thing to minimize legal risk, is not the same as minimizing cost or complexity, or doing what's best for the IETF and the community. The narrow goal of minimizing the immediate legal risk is hardly the only, or even an especially good, reason to seek advice of counsel. Counsel is there to assist you in understanding the legal implications of your possible choices and then in implementing the choice you make. They are not there to make your decisions for you. Ned
Re: [EAI] Last Call: draft-ietf-eai-popimap-downgrade-07.txt (Post-delivery Message Downgrading for Internationalized Email Messages) to Proposed Standard
The IESG has received a request from the Email Address Internationalization WG (eai) to consider the following document: - 'Post-delivery Message Downgrading for Internationalized Email Messages' draft-ietf-eai-popimap-downgrade-07.txt as Proposed Standard Rats; I missed this in earlier review: -- Section 3.2.1 -- This procedure may generate empty group elements in From:, Sender: and Reply-To: header fields. [I-D.leiba-5322upd-from-group] updates [RFC5322] to allow (empty) group elements in From:, Sender: and Reply-To: header fields. It does not. It adds group syntax to From only. Group syntax is already allowed in Reply-To, and the draft does not change the rules for Sender. If this spec also needs Sender to change, I can update the 5322upd-from-group for that, but it's not there now. I think the answer to that is yes, this needs to be allowed for Sender. Ned
Re: [EAI] Last Call: draft-ietf-eai-popimap-downgrade-07.txt (Post-delivery Message Downgrading for Internationalized Email Messages) to Proposed Standard
I think 5322upd-from-group needs to apply to both backward-pointing address types to be effective for what popimap-downgrade needs. I will make the change. But in any case, the downgrade doc needs to remove reply-to from the list of header fields that 5322upd-from-group changes. Maybe this: OLD This procedure may generate empty group elements in From:, Sender: and Reply-To: header fields. [I-D.leiba-5322upd-from-group] updates [RFC5322] to allow (empty) group elements in From:, Sender: and Reply-To: header fields. NEW This procedure may generate empty group elements in From:,Sender: and Reply-To: header fields. [I-D.leiba-5322upd-from-group] updates [RFC5322] to allow (empty) group elements in From: and Sender:. Looks like the correct change to me. Ned
Re: Draft IESG Statement on Removal of an Internet-Draft from the IETF Web Site
In the case of DMCA i am not competent to speak to circumstances surrounding a dmca. i am glad you and all the other engineers here are. sure saves the ietf lawyer a lot of work. Bingo. And even if we were competent to assess this stuff - which we most assuredly are not - any notion that a specific policy can be drafted that would cover all the eventualities that could arise in the context of a DMCA notice is preposterous on its face. Anyone who has looked at even a few DMCA situations knows they range from looneys with no grounds and no resources to entirely legitimate cases of gross infringement to what amounts to a SLAPP suit. If and when this happens the IETF will have to deal with the situation as it presents, hopefully assisted by competent legal counsel. The only question that need concern us at present is whether or not the stated policy gives the IESG the necessary flexibility. It seems to me that it does. Ned
Re: Basic ietf process question ...
On 03/08/2012, at 8:09 PM, ned+i...@mauve.mrochek.com wrote: Very much; when it becomes a document (e.g., mixed markup), XML is a much better choice. The other interesting case is where large amounts of data arrive in a stream. SAX and SAX-like libraries makes this easy to implement with XML. I hope there's an equivalent for Json; if not there needs to be. Funny you mention that, I was just looking into that yesterday. This seems to be in the front running: http://lloyd.github.com/yajl/ This looks promising. Thanks for the pointer. Ned
Re: Basic ietf process question ...
XML Schema is grossly over-engineered for its purpose; being the subject of requirements from not only the document markup field, but also databases and object models, it's a twisted monstrosity that tries to do many things, and fails at most. Agreed, and I would add that it is seriously lacking in flexibility. (I would like to have words with whoever thought the unique particle attribution rule was a good idea.) That said, there's a timing issue in play here. A variety of tools supporting other validation systems, e.g., RelaxNG or Schematron are available now, but a few years back alternatives to XML Schema were much more limited. Despite our dislike we've used XML Schema in a couple of places for precisely this reason. Specifically, it's very common for people to try to use schema to inform binding tools into specific languages. However, the underlying metamodel of XML, the Infoset, is both complex and a poor fit for most languages, so bindings take shortcuts and expose a profile of XML's range of expression, encouraging some patterns of use, while discouraging (or disallowing) others. Since the bindings often make different decisions (based upon the language of use), interoperability is difficult (sometimes, impossible). It very much depends on what you're doing and how you're doing it. If what you want is for your data to manifest directly as a data structure, XML is a lousy fit for that for a bunch of different reasons. Json is the clear choice in such cases. But there are other uses where the more complex Infoset of XML can be an asset. Really, it's all about how you use the available tools. Furthermore, designing a schema that is extensible is incredibly convoluted in XML Schema 1.0. Schema 1.1 was designed to address this failure, but it hasn't been broadly adopted; most people I know in the field consider it a failure. Yes, XML Schema makes this a lot harder to do than it should be, but in a lot of designs I've seen it also has to do with how XML is actually used. A bad design is a bad design, regardless of what schema language you use. What surprises me and many others is that people are still using it and promoting it, when it's well-understood by almost EVERYONE who was involved in using XML for protocols in the past ten years agrees that it's a mistake. See above. I certainly wouldn't use XML Schema for anything new, but there's a lot of legacy stuff out there. Ned
Re: Basic ietf process question ...
On 03/08/2012, at 5:59 PM, Ned Freed ned.fr...@mrochek.com wrote: Specifically, it's very common for people to try to use schema to inform binding tools into specific languages. However, the underlying metamodel of XML, the Infoset, is both complex and a poor fit for most languages, so bindings take shortcuts and expose a profile of XML's range of expression, encouraging some patterns of use, while discouraging (or disallowing) others. Since the bindings often make different decisions (based upon the language of use), interoperability is difficult (sometimes, impossible). It very much depends on what you're doing and how you're doing it. If what you want is for your data to manifest directly as a data structure, XML is a lousy fit for that for a bunch of different reasons. Json is the clear choice in such cases. But there are other uses where the more complex Infoset of XML can be an asset. Very much; when it becomes a document (e.g., mixed markup), XML is a much better choice. The other interesting case is where large amounts of data arrive in a stream. SAX and SAX-like libraries makes this easy to implement with XML. I hope there's an equivalent for Json; if not there needs to be. Really, it's all about how you use the available tools. Furthermore, designing a schema that is extensible is incredibly convoluted in XML Schema 1.0. Schema 1.1 was designed to address this failure, but it hasn't been broadly adopted; most people I know in the field consider it a failure. Yes, XML Schema makes this a lot harder to do than it should be, but in a lot of designs I've seen it also has to do with how XML is actually used. A bad design is a bad design, regardless of what schema language you use. What surprises me and many others is that people are still using it and promoting it, when it's well-understood by almost EVERYONE who was involved in using XML for protocols in the past ten years agrees that it's a mistake. See above. I certainly wouldn't use XML Schema for anything new, but there's a lot of legacy stuff out there. That's the rub, isn't it? Yeah, and it sure is rubbing me the wrong way every time I look at our usage. I know it was the right choice at the time, and now that it's done it's not cost effective to change unless we need additional capabilities, but that all so ... unsatisfying. Ned
Re: New Version Notification for draft-leiba-3777upd-eligibility-00.txt
I'd probably also recommend excluding paid employees of ISOC. I cannot really think of rationale that applies to the secretariat staff but not ISOC. perhaps we should take the leap of assuming folk are adults here (i realize it is a stretch), and not start a black-list with no proof of termination. +1 on all points, including the stretch part. Ned P.S. I'm not a big fan of for appearance's sake. All too often it proves to be a path to madness.
Re: registries and designated experts
It seems to me that if an expert reviewer thinks that something will do notable harm, they should decline to make a decision and defer it to the IETF at large so they are not an expert, they are a rubber stamp? bs. +1 More generally, the notion of appealing to the IETF at large, whatever that is supposed to consist of, isn't part of the written policy of any registry I'm aware of. There's supposed to be an appeals process (usually to the IESG), and that's the obvious process to invoke should an issue come up that a reviewer feels uncomfortable resolving directly. Additionally, given the widespread use of IANA registries by people and organizations not connected to the IETF in any way - and this usage is something we want to encourage - appealing registrations to the IETF at large is a really bad idea. Ned
Re: RFC 2119 terms, ALL CAPS vs lower case
I find this morning a message on the URN WG list by Alfred Hines on RFC 6329, which has a new (AFAIK) convention on normative language 3. Conventions Used in This Document The key words MUST, MUST NOT, REQUIRED, SHALL, SHALL NOT, SHOULD, SHOULD NOT, RECOMMENDED, MAY, and OPTIONAL in this document are to be interpreted as described in [RFC2119]. The lowercase forms with an initial capital Must, Must Not, Shall, Shall Not, Should, Should Not, May, and Optional in this document are to be interpreted in the sense defined in [RFC2119], but are used where the normative behavior is defined in documents published by SDOs other than the IETF. I am not sure this is in the direction of greater clarity. Should there be a need to overlay different degrees of normativeness onto a text, XML would probably be better bet. Whether the previous sentence is normative or not is left as an exercise for the reader. By my count, there are also two lower case musts and six lower case shoulds in there. In a document with compliance language this complex, those SHOULD have been eliminated. Be that as it may, this is asking more of the convention that it realistically can be expected to deliver. I don't know the circumstances behind this document - maybe there was no alternative - but the right thing IMO is to try and avoid having to do anything like this. Ned
Re: RFC 2119 terms, ALL CAPS vs lower case
I recommend an errata to RFC 2119: These words MUST NOT appear in a document in lower case. first, that is not an erratum, it is a non-trivial semantic change. And therefore explicitly disallowed by the erratum process. second, do we not already have enough problems being clear and concise without removing common words from our language? A very emphatic +1. Ned
Re: RFC 2119 terms, ALL CAPS vs lower case
The key words MUST, MUST NOT, REQUIRED, SHALL, SHALL NOT, SHOULD, SHOULD NOT, RECOMMENDED, MAY, and OPTIONAL in this document are to be interpreted as described in [RFC2119] when they appear in ALL CAPS. These words may also appear in this document in lower case as plain English words, absent their normative meanings. i like this a lot I agree. In fact I just incorporated it into the media types registration update. Ned
Re: RFC 2119 terms, ALL CAPS vs lower case
On May 16, 2012, at 5:22 PM 5/16/12, ned+i...@mauve.mrochek.com wrote: The key words MUST, MUST NOT, REQUIRED, SHALL, SHALL NOT, SHOULD, SHOULD NOT, RECOMMENDED, MAY, and OPTIONAL in this document are to be interpreted as described in [RFC2119] when they appear in ALL CAPS. These words may also appear in this document in lower case as plain English words, absent their normative meanings. i like this a lot I agree. In fact I just incorporated it into the media types registration update. To be sure of meaning and help confusion avoidance, I would prefer that the key words not appear in the document in lower case and that authors use the suggested replacement words (or break out the thesaurus?). Preferring it is one thing; I'm OK with that. Making it some sort of hard-and-fast rule is another matter entirely. We have too many of those as it is. Ned
Re: Leverage Patent Search API to reduce BCP79 related issues [was:
Russ Housley wrote: BCP 79 says: Reasonably and personally known: means something an individual knows personally or, because of the job the individual holds, would reasonably be expected to know. This wording is used to indicate that an organization cannot purposely keep an individual in the dark about patents or patent applications just to avoid the disclosure requirement. But this requirement should not be interpreted as requiring the IETF Contributor or participant (or his or her represented organization, if any) to perform a patent search to find applicable IPR. Your suggestion seems to be in direct conflict with BCP 79. IMHO your quote from BCP79 (page 4, bullet l.) is a very important point in BCP79. I can not speak for others, but the Internet Proxy of our company blocks urls to all well known online patent search and patent publication sites (and these are the only blocked sites I've ever encountered) for the simple reason of the insane US patent laws with this 3-fold punitive damages for willful infringement of a patent, i.e. a patent that was known to exist when shipping. As engineers, we MUST NOT read any patents because of this. All of the patent reasarch stuff is done by patent lawyers exclusively. I don't know if a similar block exists, but the policy is the same for us. Such a policy, if implemented, could easily lead to various people being unable to participate in the process. btw. I personally know only about one patent where I refused to be listed as inventor because I considered my contribution to it to be obvious to someone skilled in the technology. When I read the description after it had been processed by patent lawyers, I had serious difficulties understanding what that text meant... A pretty common occurance, unfortunately. Ned
Re: Last Call: draft-levine-application-gzip-02.txt (The application/zlib and application/gzip media types) to Informational RFC
I do believe that, someday, someone should try to write up an up-to-date description of the difference that recognizes the fact that compressed files are in use as media types with application/zip (in assorted spellings) and application/gzip (from this spec and in assorted spellings) as examples. But I now believe it is a separate task that should not block this document or registration. That pretty much sums up my view as well. I'll be happy to do that if I can ever find enough spare time to write it. You're right, it would be nice if there were some way to distinguish containers from content in MIME types. But given the existing historical mess, and that some kinds of compression are just a different way to encode a bunch of bits (zlib) whereas others are more like a small filesystem (zip and tgz), even if we could start with a clean sheet it's not obvious to me what would be the best thing to do. This is further complicated by the fact that there are now a number of types defined that are actually zip with specific semantics attached to the content. There are also types defined for use only within such containers. Ned
RE: Last Call: draft-levine-application-gzip-02.txt (The application/zlib and application/gzip media types) to Informational RFC
-Original Message- From: ietf-announce-boun...@ietf.org [mailto:ietf-announce-boun...@ietf.org] On Behalf Of The IESG Sent: Thursday, May 03, 2012 2:21 PM To: IETF-Announce Subject: Last Call: draft-levine-application-gzip-02.txt (The application/zlib and application/gzip media types) to Informational RFC The IESG has received a request from an individual submitter to consider the following document: - 'The application/zlib and application/gzip media types' draft-levine-application-gzip-02.txt as Informational RFC The IESG plans to make a decision in the next few weeks, and solicits final comments on this action. Please send substantive comments to the ietf@ietf.org mailing lists by 2012-05-31. Exceptionally, comments may be sent to i...@ietf.org instead. In either case, please retain the beginning of the Subject line to allow automated sorting. Only two things: 1) Shouldn't this be Standards Track? Based on past practices, the answer to that seems to be no. See, for example, RFC 6208, RFC 6207, RFC 6129, RFC 5967, and so on. The rule seems to that if the specification defines the format as well as registering one or more media types, then it needs to be on the standards track. But if the specification simply registers a bunch of types, then it's informational. And I think this makes sense. After all, how would one assess interoperability or pretty much anything else of a registration? I suppose you could argue that in cases where the registration points at an external specification (which is not the case here) that it could be a sort of hook to allow such assessment, but I think we have enough problems performing evaluations of our own work that we don't need to added fun of evaluating externally defined formats. 2) Should it mention the xdash draft, since it talks about application/x-* types? That specification is really focused on not using x- in newly defined registries, not on how it's used in existing registries. And since x-gzip and friends are only mentioned to say they shouldn't be used, I'd say that while such a reference would not bother me, I don't think it's especially helpful. Ned
RE: Last Call: draft-levine-application-gzip-02.txt (The application/zlib and application/gzip media types) to Informational RFC
-Original Message- From: Ned Freed [mailto:ned.fr...@mrochek.com] Sent: Thursday, May 03, 2012 2:54 PM To: Murray S. Kucherawy Cc: ietf@ietf.org Subject: RE: Last Call: draft-levine-application-gzip-02.txt (The application/zlib and application/gzip media types) to Informational RFC 1) Shouldn't this be Standards Track? Based on past practices, the answer to that seems to be no. See, for example, RFC 6208, RFC 6207, RFC 6129, RFC 5967, and so on. [...] 2) Should it mention the xdash draft, since it talks about application/x-* types? That specification is really focused on not using x- in newly defined registries, not on how it's used in existing registries. And since x- gzip and friends are only mentioned to say they shouldn't be used, I'd say that while such a reference would not bother me, I don't think it's especially helpful. Good enough, just thought I'd ask. I support publication as-is. As do I. Ned
RE: 'Geek' image scares women away from tech industry ? The Register
From: Mary Barnes [mary.ietf.bar...@gmail.com] Here is an article that does a far better job of explaining the situation than I did: http://www.todaysengineer.org/2011/May/women-in-engineering.asp The largest reason women leave engineering is due to the work environment and perceived lack of support from colleagues. Even the summary of the study that is given in the article is a fascinating read, with a lot of information on the subject that I've never seen before. Who would have expected: The study also found that 15 percent of women who earn undergraduate degrees in engineering never entered the profession at all. Many of them went on to enter the legal or medical professions or other fields where their engineering education served them well. The full study is at: http://www.studyofwork.com/wp-content/uploads/2011/03/NSF_Women-Full-Report-0314.pdf That link appears to be broken. The link that worked for me is: http://studyofwork.com/files/2011/03/NSF_Women-Full-Report-0314.pdf Ned
Re: Proposed IESG Statement on the Conclusion of Experiments
Ned, On Apr 25, 2012, at 7:31 PM, Ned Freed wrote: I see no value in deallocating code point spaces It depends on the size of the space. Why? Because if you deallocate and reallocate it, there can be conflicts. Perhaps you haven't noticed, but a lot of times people continue to use stuff that IETF considers to be bad ideas, including but not limited to things we called experiments at some point. Perhaps you haven't noticed, but no one was suggesting deallocating and reallocating anything that was in use. Or do you have a different interpretation of if appropriate? How can you possibly determine with any degree of reliability if something you know deployed to some extent is still in use or not? The Internet is a big place. Again, the *only* case where it makes sense to deallocate is if the space is small. In such cases the rewards outweigh the risks. And getting rid of information that people may need to get things to interoperate seems to, you know, kinda go against some of our core principles. Sorry, where did anyone suggest getting rid of any information that people may need to get things to interoperate again? Or do you interpret moving a XML page from a web server into an informational RFC to be getting rid of information? Yes I most certainly did, because that what it amounts to. The instant you move the information to a new place and break the old pointers to it, you have effectively gotten rid of it. I'll admit I find this thread bordering on the surreal with some fascinating kneejerk reactions. What's surreal is your belief that the sorts of actions you're proposing have no consequences. As far as I can tell, the only thing that was proposed was something to encourage documentation of the conclusion of experiments and if appropriate, deprecate any IANA code points allocated for the experiment. Yes, the original statement said deprecate, and I had no problem with it. But this quickly changed to people saying that code points need to be deallocated, which is what I was responding to. Here's a direct quote from an early message on this thread: From my experience at IANA, trying to figure out who to contact to remove a code point gets harder the longer the code points are not being used. Unless the code space is unlimited, I'd argue that you want to deallocate as soon as an experiment is over. remove and deallocate. Not deprecate. And no code space is unlimited. Oh, and you're the one who wrote this. Both of these seem like good things to me. This has somehow been translated into variously: a) a declaration about how research is done b) deletion and/or reallocation of code point spaces that people are using c) killing off successful protocols because they're documented in experimental not standards track rfcs d) violating our core principles e) process for the sake of process f) IANA being a control point for the Internet g) etc. Did I miss a follow-up message from the Inherently Evil Steering Group that proposed these sorts of things? Did I ever say that I was repsonding to the original IESG statement? No, I don't think I ever said that. Anyway, I've made my point, and as PHB said, you've now devolved to stupid tricks to bolster your argument. I'm done. Ned
Re: Proposed IESG Statement on the Conclusion of Experiments
I see no value in deallocating code point spaces and a huge amount of potential harm. It depends on the size of the space. I completely agree that if the space is large - and that's almost always the case - then deallocating is going to be somewhere between silly and very damaging. Deprecacting the use of a code point, OTOH, may make sense even if the space is large. The takeaway here, I think, is that if you're going to conclude experiments, you need to examine these allocations and do something sensible with them, where sensible is rarely going to mean deallocate. Except at the very lowest levels of the protocol stack (IP and BGP) there is really no technical need for a namespace that is limited. We do have some protocols that will come to a crisis some day but there are plenty of ways that the code space for DNS, TLS etc can be expanded if we ever need to. There may not be any technical need, but there are a number of legacy designs that were done ... poorly. The Internet is driven by innovation and experiment. There are certainly people who think that the role of this organization is to protect the Internet from the meddling of people who might not get it but they are wrong. +1 Even more wrong is the idea that IANA can actually act as suggested. IANA only exists by mutual consent. I am happy to register a code point if there is a reasonable procedure to do so. But if the idea is to force me to kiss someone's ring then I'll pick a number and ship the code and let other folk work out the problems later. This already happens on a significant scale in the wild. SRV code points being a prime example. There are far more unofficial code points than recognized ones. Some of them are in standards I wrote at W3C and OASIS. It would be best if IANA tracked these but I think it rather unlikely anyone is going to accidentally overwrite the SAML or the XKMS SRV code points. Media types are an excellent example of this. The original registration procedures were too restrictive so people simply picked names to use. We fixed that for vendor assignments (fill in a web form) and the registrations starting rolling in. (We're now trying to do the same for standard assignments.) But of course we now have a legacy of unassigned material to deal with. It has happened with other registries too. Back in the day there was an attempt to stop the storage manufacturers using ethernet IDs on their drives. So the drive manufacturers simply declared a block of space (Cxx) as theirs and the IEEE has since been forced to accept this as a fait accompli. It has happened in UPC codes as well, when the Europeans decided that the US authority was charging them ridiculous fees they just extended the code by an octet and set themselves up as a meta-registry. The only role of IANA is to help people avoid treading on each other by accident. If it starts issuing code points that have been 'lightly used', the value of IANA is degraded. I certainly don't want my protocol to have to deal with issues that are caused by someone's 'experiment' that has not completely shut down. The only value of going to IANA rather than choosing one myself is to avoid contamination with earlier efforts. No, not quite. The other important role is (hopefully) to keep people from having multiple allocations for the same thing. The only area where I can see the benefits of re-allocation outweighing the risks is in IP port assignments but even there I think the real solution has to be to simply ditch the whole notion of port numbers and use SRV type approaches for new protocols. +1 IANA is not a control point for the Internet. A fact that people need to keep in mind when the ITU attempts to grab control in Dubai later on this year. Weakness is strength: The registries are not the control points people imagine because they only have authority as long as people consent. Good point. Ned
Re: Proposed IESG Statement on the Conclusion of Experiments
Ned, On Apr 25, 2012, at 10:46 AM, Ned Freed wrote: I see no value in deallocating code point spaces and a huge amount of potential harm. It depends on the size of the space. Why? Because if you deallocate and reallocate it, there can be conflicts. Perhaps you haven't noticed, but a lot of times people continue to use stuff that IETF considers to be bad ideas, including but not limited to things we called experiments at some point. We're talking about completed experiments. It doesn't matter if we're talking about pink prancing ponies. The issue is whether or not the code point is valuable. If it isn't there's no reason to deallocate it and every reason not to, although you may want to deprecate it's use if the experiment proved to be a really bad idea. I'm unclear I see any particular value in having IANA staff continue to maintain registries (which is what I've translated code point _spaces_ to) for the protocol defined within the RFC(s) of that experiment. Ah, I think I see the conflict. You're talking about experiments that define namespaces themselves, rather than using code points out of some other more general space, which is what I was talking about. That said, I have to say I reach pretty much the same conclusion for experimental code point spaces that I do for experimental use of code points in other spaces: They should not be deallocated/deleted. Again, just because the IETF deems an experiment to be over, or even if the IETF concludes that the experiement was a failure, doesn't mean people won't continue to use it. And getting rid of information that people may need to get things to interoperate seems to, you know, kinda go against some of our core principles. I could, perhaps, see memorializing the final state of the registries for the experiment in an informational RFC, but don't really see the point in cluttering up http://www.iana.org/protocols with more junk than is already in there. Trying to find things is already annoying enough. That's a problem with the organization of the web site, not an argument for getting rid of information. There are abundant examples of web sites that contains thousands of times as much stuff as IANA does where finding what you want is easy if not outright trivial. Ned
Re: Issues relating to managing a mailing list...
Is this really a big enough problem to be worth solving? I can't recall a single instance where I received IETF list with a problematic attachment. i travel to places with very poor bandwidth. it is a problem. and the vast majority of users just do not get it. we send 20mb documents around the office all the time. suck it up. for who has problem in attachment downloading the solution should be at the delivery Message Store level, where the strip of the attachment could be done accordingly to an user configurable mailbox parameter (as we do on our server, where we call it Easy Delivery). Or better still, don't strip anything. Use a reasonably capable client and that doesn't fetch attachments unless you tell it to. I'll again note that implementation of this policy will cause people to send single part messages in order to avoid attachment stripping, which will make it impossible for clients to avoid downloading material they do not want. In other words, it will exacerbate, rather than solve, the problem. Ned
RE: Issues relating to managing a mailing list...
--On Thursday, March 15, 2012 00:00 -0400 Ross Callon rcal...@juniper.net wrote: I don't like this proposal for two reasons: I frequently read email while not connected; When connected, bandwidths have gotten high enough that attachments on the most part are not slowing things down in an uncomfortable way. It might be okay for really large attachments, as long as only a few messages are affected. Borrowing a bit from Randy, the solution to really large attachments is to ban them. Personally, I'd find it perfectly reasonable to have any message in the megabyte range or above (or probably even an order of magnitude smaller) rejected with a message that amounted to if you have that much to say, write an I-D, post it, and point to it. That is much more plausible today, when the mean time between I-D submission and posting is measured in minutes (except during blackout periods) than when it was in days. During blackout periods, the last thing the community needs is people adding to already-overloaded lists by posting long documents in email. If people want to use up part of their maximum size quota by posting html in addition to text, or appending long disclaimers or autobiographies, that shouldn't the community's problem. You begin by talking about banning large attachments. You then segue into a discussion where you talk about a maximum size that includes the primary message content, not attachments, then you throw in disclaimers, which may or may not be attachments. Other have followed up by supporting the limit on attachment size, others still have talked about banning attachments regardless of size. Do you see the problem here? The minute you start focusing on specifics of message content, you're in a rathole. What counts as an attachment? (And yes, we have a precise definition for what constitutes an attachement, but following that definition gives people the ability to route around it.) It follows that any limit needs to be on overall message size. (Even this is a little perilous because message sizes can change due to MIME downgrading or upgrading, both of which happen regularly.) I would not be opposed to imposing such a limit, although it's going to need to be higher than some people would probably like - it's surprising how easily you can approach 1Mb with a single part, plain text message containing nothing remotely resembling an attachment. Ned
Re: Issues relating to managing a mailing list...
Is this really a big enough problem to be worth solving? I can't recall a single instance where I received IETF list with a problematic attachment. i travel to places with very poor bandwidth. it is a problem. and the vast majority of users just do not get it. we send 20mb documents around the office all the time. suck it up. for who has problem in attachment downloading the solution should be at the delivery Message Store level, where the strip of the attachment could be done accordingly to an user configurable mailbox parameter (as we do on our server, where we call it Easy Delivery). Or better still, don't strip anything. Use a reasonably capable client and that doesn't fetch attachments unless you tell it to. Yes, I agree, but we can leave the end user to choose accordingly their own no capable/capable clients or maybe accordingly to temporary condition. You can as long as attachments remain as attachments, which will no longer be true if this proposal is implemented. If they don't the strip attachments approach no longer works. You either (a) Only strip attachments, which leaves you with large messages to deal with, (b) Strip by part size, which will leave you with empty message, or (c) Cut parts down to a given size, which is guaranteed to work poorly. The two list approach, if secondary list proves popaular, may lead to the same result, BTW. As I commented previously in a private email, it never ceases to amaze me how people in this community only apply the route around problems mantra when it suits their purpose to do so. Ned
RE: Issues relating to managing a mailing list...
John, I agree completely with everything you say here. Ned --On Thursday, March 15, 2012 08:16 -0700 Ned Freed ned.fr...@mrochek.com wrote: ... It might be okay for really large attachments, as long as only a few messages are affected. Borrowing a bit from Randy, the solution to really large attachments is to ban them. Personally, I'd find it perfectly reasonable to have any message in the megabyte range or above (or probably even an order of magnitude smaller) rejected with a message that amounted to if you have that much to say, write an I-D, post it, and point to it. That is much more plausible today, when the mean time between I-D submission and posting is measured in minutes (except during blackout periods) than when it was in days. During blackout periods, the last thing the community needs is people adding to already-overloaded lists by posting long documents in email. ... You begin by talking about banning large attachments. You then segue into a discussion where you talk about a maximum size that includes the primary message content, not attachments, then you throw in disclaimers, which may or may not be attachments. Other have followed up by supporting the limit on attachment size, others still have talked about banning attachments regardless of size. Do you see the problem here? The minute you start focusing on specifics of message content, you're in a rathole. What counts as an attachment? (And yes, we have a precise definition for what constitutes an attachement, but following that definition gives people the ability to route around it.) Ok, that is fair. At best, I skipped being explicit about several steps in my thinking. I personally see the attachment issue as almost irrelevant, in part because of the what constitutes an attachment issue you identify above. In partial defense, I was distracted a bit by comments from others about the first N bytes of a message and text and attachment models (which, as you know better than almost anyone, is a user construct about how messages might be assembled but one that doesn't necessarily map directly and usefully back from MIME body parts). I think the issue is almost entirely about (i) size, independent of how a message is structured, and (ii) what, in IMAP terms, is automatic synchronization into offline or disconnected mode. More generally, it is ability to read and work with IETF mailing lists at a time when one has little or no connectivity. From that point of view, Russ's original question is about something that is not only a bad idea but an ill-formed proposition... at least unless we could agree on and enforce a particular model of IETF participants assemble messages and send them -- a goal that I believe is so hopeless as to not be worth discussing. It follows that any limit needs to be on overall message size. (Even this is a little perilous because message sizes can change due to MIME downgrading or upgrading, both of which happen regularly.) Exactly. And that also identifies the other distraction I was trying to get at. In addition to MIME downgrading and upgrading, the size of a message can be distorted by such things as addition (possibly by submission servers or list exploders and out of user control) of lengthy explanatory headers (including some called for by IETF standards like DKIM), body text (including the notorious unless you agree to this conditions, please un-read the message announcements but all list-added footers), or duplicate content (e.g., sending both text/plain and text/html with multipart/alternative). If we were to start trying to split hairs on how to count message sizes, I think we would rapidly go down a slightly different rat hole. I would not be opposed to imposing such a limit, although it's going to need to be higher than some people would probably like - it's surprising how easily you can approach 1Mb with a single part, plain text message containing nothing remotely resembling an attachment. Yes. But an entirely anecdotal survey of recent postings to the IETF list and comparison with Thomas Narten's statistics indicates that a posting to that list of 50K in aggregate size is well out on the upper tail. Most messages of twice that size and larger are either duplicate content or inclusion of a great deal of content from multiple previous messages. I'd be happy discouraging both of those patterns for other reasons so wouldn't be upset if a size rule had the accidental side-effect of pushing back on them. I can well imagine the averages for some WG lists being much larger, especially if editors are including blocks of existing and proposed text for comment. But, as a notorious author of long messages and someone who often prefers consolidated responses to a lot of small messages, I'd
Re: Issues relating to managing a mailing list...
Some suggestions have been made about the IETF mail lists. There is a way for mailman to strip attachments and put them in a place for downloading with a web browser. This would be a significant change to current practice, so the community needs to consider this potential policy change. What do you think? Is this really a big enough problem to be worth solving? I can't recall a single instance where I received IETF list with a problematic attachment. OTOH, I routinely get IETF messages with useful attachements - typically a critical revision to a draft which for whatever reason can't be posted as an I-D - that I really need to be able see without having to bother with some indirection. There are also times when I have access to email but no web access. What am I supposed to do then? This also creates significant problems for list archiving. Right now I can easily capture the entire content of the list and preserve it for however long I want. Do this and that becomes significantly more difficult - I have to detect these indirections, fetch and stuff the attachments back in, because regardless of stated policy I cannot assume that web link will be valid for as long as I need it to be. I also worry about the ability of mailman to actually know what an attachment is versus a message constructed in multiple parts. And there's also the issue of what sort of indirection it uses. Past experience with advanced list manager features like this doesn't give me a whole lot of confidence in their MIME chops. I suppose I could live with this - but not actively support it - if the stripping was limited to abusively large attachments - say ones over 5Mb or thereabouts. But otherwise it's a TERRIBLE idea, and will simply result in everyone including the draft or whatever in the primary message text in order to avoid this nonsense, which results in a degradation of list quality for all concerned. Ned
Re: provisioning software, was DNS RRTYPEs, the difficulty with
There are some false equivalences floating around here. I don't think anyone is suggesting that having provisioning systems or even DNS servers themselves check for syntax errors in the contents of complex records like DKIM, SPF, DMARC, or whatever is necessarily a bad idea. (Whether or not it will actually happen is another matter; I'm dubious.) Rather, the issue is with requiring it to happen in order to deploy a new RRTYPE of this sort, which is the result you get if the DNS server returns some series of tokens instead of the original text string. That's the sort of thing that forces people to upgrade, or search around for a script to do the conversion (which won't even occur to some), and that's an extra burden we don't need to impose. It would still be possible to work around the need for a plugin, e.g. by depending on some wizard web site, as in John's thought experiment. For the rest of us, the possibility to install a plugin that takes care of all the nitty-gritty details, instead of having to wait for the release and distribution of the next version of BIND, can make the difference between deploying a new RR type right away and procrastinating endlessly. You're still not separating the two cases. Again, an *optional* plugin to check syntax of a record but not produce any sort of tokenized result is fine, a plugin that's *mandatory* to deploy is going to be almost as much of an impediment to deployment as requiring an upgrade. Code is code, and people don't install new code willy-nilly. The issue is to upgrade once rather than on each new RR type. Exactly. That's why mandatory plugins are a bad idea. Correct, but when you publish a complex record you are calling forth that complexity. I don't see much difference if the bug is at mines or at the remote site, since their effects are comparable. They most certainly are not. A bug in my client only affects me, a bug in the server can easily kill the entire zone. And even if separation techniques are employed, if the plugin fails the best you're going to be able to do is server out a domain with missing entries. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: provisioning software, was DNS RRTYPEs, the difficulty with
On 07/Mar/12 09:42, ned+i...@mauve.mrochek.com wrote: It would still be possible to work around the need for a plugin, e.g. by depending on some wizard web site, as in John's thought experiment. For the rest of us, the possibility to install a plugin that takes care of all the nitty-gritty details, instead of having to wait for the release and distribution of the next version of BIND, can make the difference between deploying a new RR type right away and procrastinating endlessly. You're still not separating the two cases. I think I did, however badly. Let me try by example, assuming my server doesn't recognize SPF. As it's unrecognized, I have to write it as such, e.g. TYPE99 \# 12 0B 763D73706631 20 2D613131 OK, now I'm completely confused. What does lack of SPF record support in your server input format have to to with syntax checking of the *content* of the SPF record? I can write the record as text or in hex or base64 or whatever format I want; the issue is looking *inside* the data and either (a) just checking it's syntax versus (b) checking it and turning it into some kind of tokenized stuff the DNS server actually serves out. Again, an *optional* plugin to check syntax of a record but not produce any sort of tokenized result is fine, Now the plugin can check the syntax and spot the error. I may correct it or not. If I don't, I'll just serve bad data. In any case, my zone source still contains the line above. Thus the plugin is optional. a plugin that's *mandatory* to deploy is going to be almost as much of an impediment to deployment as requiring an upgrade. Code is code, and people don't install new code willy-nilly. Any plugin that's necessary to transform a nontrivial input format into a tokenized result is effectively mandatory. Sure, you can substitute a preprocessing script that does the transform and spits out hex or whatever, but no matter how you do it there's an essential translation component involved in provisioning those records. You may be able to avoid having that component cause the entire server fail, but it's still in the critical path for setting up those records. Possibly, I can also run, say: echo 'SPF v=spf1 -a11' | plugin --dump-hex --silent and have it dump the TYPE99 line above (without signalling the error, since I said --silent). Then I copy its output and paste it into the zone source. Let's please stop talking about what you can do manually. This isn't about people whose main provisioning tool is emacs or bind, and who operate their own servers with full autonomy and report to nobody. This audience doesn't have issues with any of this stuff. Finally, I can enable automatic invocation of the plugin. That way, I can write the SPF record directly in the zone source and have my DNS-fictional server do the copy and paste on the fly for me. I wouldn't call such thing mandatory. Again, it depends on whether or not it's in the critical provisioning path. A syntax checker isn't, a parser and tokenizer is. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: provisioning software, was DNS RRTYPEs, the difficulty with
Input -P- DNS server -D- DNS stub -Q- Output P is the provisioning I think you want to break that into the provisioning interface and the data format it produces that the DNS server consumes. (My reason for that is we have a specification for at least such format, with all that implies.) I was also going to mention that. There's a lot of different formats for zone file data, with BIND-ish master files being only one of them. We seem to believe that the D part is deployed so that adding new unknown RRTypes is not an issue. I think this is correct, but OTOH someone recently asked about possible issues in this area, and if I remember correctly, received no response. Last month I ran into a guy on the dmarc list who complained that his server returns NOTIMP in response to queries for SPF records (because it doesn't implement them) and clients were doing odd things. But it's been a long time since I've run into anyone else like that, so I agree, it's not an issue. Problem is then in P and Q. I personally don't see a big problem with Q, but others clearly do so we need to leave it in. I'm not aware of problems, but I don't use Windows which is where the problems are supposed to be. Exactly - neither do I. (Well, OK, I have a Windows laptop at work to handle the company HR stuff that won't work in anything other than IE, but I try really, really hard never to use it.) Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: Last Call: draft-melnikov-smtp-priority-07.txt (Simple Mail Transfer Protocol extension for Message Transfer Priorities) to Proposed Standard
That said, I think an important omission in this document is that it only allows MSA's to change message priorities to conform to site policy. MTAs should also be allowed to do this. Can you elaborate a bit more on possible use cases? Nobody is going to simply allow priorities to be set willy-nilly on mail coming from random sources outside their administrative domain. That's absurd. However, they may be will to make bilateral arrangements with selected partners, identified by IP address or whatever, that would allow such a setting, perhaps within a restricted range of allowed values. Would such an MTA only lower the priority or do you think it might also raise it? I don't see any reason to limit it to lowering it. Another issue is the silly prohibition against using Priority: and other header fields to set priority levels. What if is existing site policy is in fact to use those fields to determine message priority? (Ignoring the question of whether use of MT-Priority header field is a good thing or not for a moment.) I actually don't have a strong feeling against usage of other existing header fields. Some of the existing header fields don't have the exact semantics desired here. Well, sure. You most definitely don't want to mix in Importance or other MUA level priority fields. Others (like the MIXER's Priority) have the right semantics but don't support sufficient number of priorities required by MMHS (6 levels). I think you're going to have to accept the fact that the overwhelming majority of people out there running email systems have never even heard of MMHS and even if they have don't give a damn about faithfully retaining it's semantics. But they do care that new mechanism be made compatible with whatever ad-hoc scheme they are currently using, even if said scheme doesn't have the full range of values. For example, I can easily see a site wanting to map this to and from the field used by Microsoft Exchange (sorry, I forget the exact name) even though if memory serves that field only accepts three values. And either this is going to happen no matter what the specification says, or the specification simply won't deploy in any meaningful way. That is why a new header field was introduced. But anyway, I am happy for this restriction to be removed/relaxed. Can you recommend some specific text? I'll try to do so later this week. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: Last Call: draft-melnikov-smtp-priority-07.txt (Simple Mail Transfer Protocol extension for Message Transfer Priorities) to Proposed Standard
Ned: Another issue is the silly prohibition against using Priority: and other header fields to set priority levels. What if is existing site policy is in fact to use those fields to determine message priority? Alexey: I actually don't have a strong feeling against usage of other existing header fields. Some of the existing header fields don't have the exact semantics desired here. Others (like the MIXER's Priority) have the right semantics but don't support sufficient number of priorities required by MMHS (6 levels). That is why a new header field was introduced. Right, this is the issue I have with Ned's desire to allow use of other fields: those fields have inconsistent semantics, and inconsistent and non-standard usage. They're often used to indicate visual importance of the message to the MUA, rather than anything related to transmission priority. I'm sorry, but that's simply a strawman. AFAICT nobody pushing back on this is suggesting mapping MTA priority to or from a field with totally different semantics. But there are plenty of fields in common use that are used to control MTA priority. Those are the fields I'm talking about. And yes, people will make mistakes. They always do. But as I commented in a previous discussion, if we let that stop us we'd never standarize anything. That said, I'd have no problem with some sort of MAY-level permission to *MSAs* to use these fields. I'd feel very uncomfortable allowing *MTAs* to do it. Ned, would it be adequate to your needs to handle it that way? No. Like it or not, if you expect this to gain any traction at all, you're going to have to accept the fact that people are going to want to tunnel it through fields with the same general semantics but restricted as to possible values. Something like this: OLD Other message header fields, such as Importance [RFC2156], Priority [RFC2156] and X-Priority, MUST NOT be used for determining the priority under this Priority Message Handling SMTP extension. I hadn't noticed importance listed as a possible source of MTA priority information. That's actually committing the error of mapping a field with entirely different semantics into MTA priority and at a minimum needs to be removed - we should NOT be encouraging that. NEW Other message header fields, such as Importance [RFC2156], Priority [RFC2156] and X-Priority, are used inconsistently and often with different semantics from MT-Priority. Message Submission Agents [RFC6409] MAY use such fields to assign an initial priority in the absence of an SMTP PRIORITY parameter. Otherwise, such fields MUST NOT be used for determining the priority under this Priority Message Handling SMTP extension. It seems you're complaining about other people doing something they never wanted to do while actually making that error yourself. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: Last Call: draft-melnikov-smtp-priority-07.txt (Simple Mail Transfer Protocol extension for Message Transfer Priorities) to Proposed Standard
2. I, too, noticed all the lower-case should and may words. I suggest that the best way to handle this is to make the following change to the RFC 2119 citation text at the beginning of section 2: NEW The key words MUST, MUST NOT, REQUIRED, SHALL, SHALL NOT, SHOULD, SHOULD NOT, RECOMMENDED, MAY, and OPTIONAL in this document are to be interpreted as described in [RFC2119] when they appear in ALL CAPS. These words also appear in this document in lower case as plain English words, absent their normative meanings. I don't mind changing that, but ID-nits gives a warning when the text about keywords is changed and *everybody* likes to complain about that. Please talk to IESG about whether using a variant of the standard text is acceptable. I've used text like this before, and the IESG has never objected to it. One advantage with this formulation, which uses the standard 2119 text and *appends* to it, is that idnits likes it and doesn't generate a warning. More generally, ID-nits is supposed to be helpful. Straightjackets are not helpful. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: Last Call: draft-melnikov-smtp-priority-07.txt (Simple Mail Transfer Protocol extension for Message Transfer Priorities) to Proposed Standard
I do prefer the latter as well (and yes, happy to remove the restriction), but I don't feel very comfortable pretending that tunneling wouldn't happen. Of course people will tunnel stuff. But will they all tunnel it the same way, in which case a standard could be useful, or will they each have their own hacks? Given the vagueness about how you can tell that something should come out of the tunnel back into the envelope, it sounds more like the latter. If the only issue was tunneling, I'm not sure it would matter as much as it does. After all, if we have a spec that says use MT-Priority and someone uses something else within their ADMD, who will even know? But this isn't just about tunneling. There are a bunch of ad-hoc priority mechanisms out there, some in fairly widespread use. If we cannot align ourselves with those nobody is going to be interested in deploying this extension. Instead they will be asking us to retain our present capability of interoperating with those existing systems. And doing both is ... messy to say the least, not to mention a violation of the current language in the specification. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: Last Call: draft-melnikov-smtp-priority-07.txt (Simple Mail Transfer Protocol extension for Message Transfer Priorities) to Proposed Standard
Other message header fields, such as Importance [RFC2156], Priority [RFC2156] and X-Priority, are used inconsistently and often with different semantics from MT-Priority. Message Submission Agents [RFC6409] MAY use such fields to assign an initial priority in the absence of an SMTP PRIORITY parameter. Otherwise, such fields MUST NOT be used for determining the priority under this Priority Message Handling SMTP extension. It seems you're complaining about other people doing something they never wanted to do while actually making that error yourself. I have no idea what you mean, so perhaps you can be more specific about what it is that you want to do. I'm talking about the inclusion of Importance: in both the original and revised text as a field we're suggesting might make sense to map. It should not be there - the semantics are not compatible with MTA priority. And even if you think it makes sense that messages marked important in the MUA should receive an MTA priority boost, allow me to point out that by GOSIP rules (which are still in effect in some places) elevated MTA priority is used to shorten the time messages are retained before returning them as undeliverable. That's really not a good idea to do for all messages marked important, and some people with systems that do this are probably unaware of the behavior. Mind you, I have no problem with someone doing it because it makes sense in their case. But we should not be suggesting doing it in the text, and getting into a discussion of the pitfalls of doing it doesn't is unnecessarily distracting. The semantics of X-Priority are less clear. It's sometimes used for MTA priority setting, other times it's been an MUA field. AFAIK it has never had the problematic semantics I just discussed. Mapping it is therefore a judgement call implementations should be allowed to make, and that makes it a good example to include in this sort of text. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: provisioning software, was DNS RRTYPEs, the difficulty with
On 05/Mar/12 18:09, John Levine wrote: Sometimes an ASCII text record will be fine, in other cases, it probably won't. My point is as we move again towards multiple text representations of the digit five for example, both encoding and parsing is easier and more secure if that digit is really for example eight bits and not text that someone has to parse. Unless you provision your DNS zones with a hex debugger, the digit will always start out as text that someone has to parse. The question is who does the parsing, the DNS server or the application. As I said in a previous message, I can see plausible reasons to put the parser into the application. Would you really want to build an SPF or DKIM parser into every DNS server? That's a lot of code that the DNS manager doesn't care about, but the mail manager does. But it would be the same code, most likely by the same author(s). It may be generic for a kind of syntax or specific for a RR type, according to its author's convenience. On a system that allows new RR types without recompiling, the code would come as some sort of plugin in both cases. There are some false equivalences floating around here. I don't think anyone is suggesting that having provisioning systems or even DNS servers themselves check for syntax errors in the contents of complex records like DKIM, SPF, DMARC, or whatever is necessarily a bad idea. (Whether or not it will actually happen is another matter; I'm dubious.) Rather, the issue is with requiring it to happen in order to deploy a new RRTYPE of this sort, which is the result you get if the DNS server returns some series of tokens instead of the original text string. That's the sort of thing that forces people to upgrade, or search around for a script to do the conversion (which won't even occur to some), and that's an extra burden we don't need to impose. Why is it important what the DNS manager cares about? Speaking as a DNS manager myself, I care a lot about being forced to upgrade. Upgrades bring unwanted bugs in other areas. In fact I'm not entirely thrilled with the idea of plugins to do some extra syntax. More code means more possibilities of bugs. I'd actually prefer to see more cross-checking of existing stuff - less code and greater benefit. Parsers, including null parsers, would come with the same sub-package that enables the new RR type definition. Their complexity would only matter to the people who read/maintain their sources. I'm sorry, but you're being naive about this. Complexity does matter to the people who just use software because added complexity translates to more bugs. PS: For anyone who didn't read my previous message, I am NOT saying that it's fine to overload everything into TXT. I am saying that new RRTYPEs that are text blobs interpreted by client software wouldn't necessarily be bad. Agreed. That doesn't preclude syntax checking on loading the zone, though. As long as we stick with syntax checking I'm (mostly) OK with it. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: provisioning software, was DNS RRTYPEs, the difficulty with
I think we should look a bit on the flow of data. If I simplify we have a flow like this: Looking at data flows is usually a good idea. Input -P- DNS server -D- DNS stub -Q- Output P is the provisioning I think you want to break that into the provisioning interface and the data format it produces that the DNS server consumes. (My reason for that is we have a specification for at least such format, with all that implies.) D is the DNS protocol Q is the query/parsing in the consumer of the data What we want is (as described in the IAB RFC on extensions of DNS) the ability to query (in D) for as precise data as possible in the triple {owner, type, class}. Some RR types like NAPTR and TXT have flaws where the selection between records in an RRSet is not in this triple. In some applications that is resolved by having a prefix to the owner. In some other applications that is resolved by parsing the RRSet. We all do believe that IF it was easier to add a new RRType for each application that would be an architecturally better solution (as adding prefix to the owner have its drawbacks). Now, the question is what blocks the ability to add new RRTypes. Yes, I think we have agreement on all of that. We seem to believe that the D part is deployed so that adding new unknown RRTypes is not an issue. I think this is correct, but OTOH someone recently asked about possible issues in this area, and if I remember correctly, received no response. Problem is then in P and Q. I personally don't see a big problem with Q, but others clearly do so we need to leave it in. And when we talk about parsing, we talk about what the mapping between provisioning and DNS packet format is. I think we need to be a little finer grained than that, per the above. Are we aligned so far? Yes, pretty much. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: provisioning software, was DNS RRTYPEs, the difficulty with
On 03/Mar/12 00:13, John R. Levine wrote: Until provisoning systems accept UNKNOWN record types they will always be a bottle neck. Without that you will have to wait for the change request to be processed. Given the history just getting records added to most of these system it will be forever. was unusually painful, since it requires adding a parser for IPv6 addresses. (Having hacked it into my provisioning system, I speak from experience.) Most new RR types are just strings, numbers, names, and the occasional bit field. Yeah, and if ISPs already had troubles with ho-de-ho-de-ho-de-ho, how will they join in on skee-bop-de-google-eet-skee-bop-de-goat? Given that, designers of new RR types will want to stick to string formats just to spare ISPs some parsing, at the cost of losing a half of the advantages that RFC 5507 talks about, along with syntactic validations aimed at preventing some permerror/permfail cases. Doubtful. If a record needs to have, say, a priority field, or a port number, given the existence of MX, SRV, and various other RRs it's going to be very difficult for the designers of said field to argue that that should be done as ASCII text that has to be parsed out to use. More generally, this is trying to have things both ways. We can't simultaneously assert that deploying simple new RRs is a breeze, making this unnnecessary, and that it's so difficult that everything should be crammed into TXT Format no matter the actual structure is. NNed ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: provisioning software, was DNS RRTYPEs, the difficulty with
On 3 mar 2012, at 16:56, ned+i...@mauve.mrochek.com wrote: Doubtful. If a record needs to have, say, a priority field, or a port number, given the existence of MX, SRV, and various other RRs it's going to be very difficult for the designers of said field to argue that that should be done as ASCII text that has to be parsed out to use. Agree with you but too many people today just program in perl och python where the parsing is just a cast or similar, and they do not understand this argument of yours -- which I once again completely stand behind myself. Regardless of the language you program in, you still have to get your proposal approved. For that to happen the folks who review these things would in effect be conceding that there is in fact a major deployment problem out there. That's the point I was trying to make, not that people wouldn't argue for this approach. I am unaware of any counterexamples that have actually made it through the process, but if they exist feel free to point them out. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: Last Call: draft-melnikov-smtp-priority-07.txt (Simple Mail Transfer Protocol extension for Message Transfer Priorities) to Proposed Standard
This draft also defines the MT-Priority header field. �It is quite unusual for a SMTP extension specification to define a mail header field. ... This is my major concern about this protocol as well, as I note in the PROTO writeup (which, unfortunately, can't be seen publicly because of a limitation in the datatracker; perhaps I should post it here). I'm interested in hearing whether others share this concern, and what the community consensus is about it. I have no problem logging stuff from the SMTP session into a message header, we've been doing that since forever. But I have a lot of problem turning a header back into SMTP for a subsequent relay, for two reasons. One is just that it's ugly. There were good reasons to separate the envelope from the body back in the 1970s, and I think those reasons are just as good now. Over in the EAI group, there was a valiant effort to tunnel EAI messages through non-EAI SMTP servers via downgrades, and we eventually gave up. Now, if you want to send an EAI message, there has to be an EAI path all the way from the sender to the recipient. This isn't exactly analogous, but it's instructive. The other reason is that I think it's too ill-specified to interoperate. The theory seems to be that the relay server notices an MT-Priority: header, and (vigorous waving of hands) somehow decides if the message has sufficient virtue to promote the header back into the SMTP session. The hand-waving is not persuasive; the suggestion of S/MIME, which obviously won't work, is not encouraging. Both good points. While I personally am indifferent to header modification being an issue - I believe we've long since crossed that bridge - I will also note that simply changing this proposal to always insert the MT-Priority: field at submission time and leave it unchanged throughout message transit would leave most of the functionality intact and eliminate the header modification issue entirely. But the upleveling of header to envelope would remain. John is quite correct in his assertion that these are mostly uncharted waters: The only remotely comparable case I can think of is the mapping of some X.400 fields to the NOTARY extension, but in that case the mapping was from X.400 envelope material to SMTP envelope material. Additionally, the mapping was precisly specified with no wiggle room for site policy. I think this will likely lead to interoperability problems. Now, in the case of message processing priority the chance of such problems causing significant trouble seem pretty slim, but it doesn't work well but it also doesn't matter much is not a justification I feel comfortable embracing. If you want to brave the ugliness and standardize this, the spec needs to explain how the server decides whether to believe an MT-Priority header. If it doesn't, it's too vague to interoperate. So I have two suggestions. One is to leave it as is, and make it experimental. If it turns out the tunnels all work the same way, you can come back and add the spec about how they work and rev it as standards track. The other is to take out the tunnel bit and make it standards track now, so a compliant implementation needs priority-aware MTAs from end to end. Even if you take out the tunnel stuff in the spec, you can still tunnel in your implementations, using whatever non-standard ad hoc kludges you want. Since the current spec has gaps that would need to be filled by non-standard ad hoc kludges anyway, you haven't lost anything. I have to say I strongly prefer the latter, especially if it also includes eliminating the prohibition on using other sources of priority information to set the envelope priority level on submission. In fact given that prohibition I see little point in implementing this, at least not in a fashion that conforms to the specification. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: provisioning software, was DNS RRTYPEs, the difficulty with
I've rarely read a message on this list that I agreed with more. +1000 to everything said below. Ned By the way, what's your opinion of draft-levine-dnsextlang-02? It isn't clear to me what problem you're trying to solve. For resolving name servers 3597 should be enough. For authoritative name servers your idea is sort of interesting, but generally upgrading the software to pick up support for new RRtypes is a good idea, since you'll also pick up new security fixes, etc. Oh, wow. Now I see the problem: people here are totally unaware of what using DNS software is like if you're not a DNS weenie. If you think that it's reasonable to require a new version of your DNS software every time there's a new RR, you've conceded total defeat. On most servers I know, software upgrades are risky and infrequent. It could easily take a year for a new version of BIND to percolate out of the lab, into the various distribution channels, and to be installed on people's servers, by which time everyone has built their applications with TXT records and it's too late. Moreover, the servers aren't the hard part, it's the provisioning systems. You and I may edit our zone files with emacs, but civilians use web based things with pulldown menus and checkboxes. If they can't enter an RR into one of them it doesn't matter whether the name server supports it or not. This latest round of teeth gnashing was provoked by discussions in the spfbis WG, where we're wondering whether to deprecate the type 99 SPF record. Among the reasons it's unused in practice is that nobody's provisioning system lets you create SPF records. The point of my draft is that it's an extension language that both name servers and provisioning systems can read on the fly. Add a new stanza to /etc/rrtypes and you're done, no software upgrades needed. The risk is much lower since the worst that will happen if the new stanza is broken is that the provisioning software and name servers will ignore it, not that the whole thing will crash. Sure, there are some RRs with complex semantics that can't be done without new code (DNSSEC being the major example), but most RRs are syntactically and semantically trivial, just a bunch of fields that the server doesn't even interpret once it's parsed the master records or their equivalent. Until the DNS crowd admits that provisioning systems are a major bottleneck, and telling people to hand-code more types into them isn't going to happen, there's no chance of getting more RRs deployed. Regards, John Levine, jo...@iecc.com, Primary Perpetrator of The Internet for Dummies, Please consider the environment before reading this e-mail. http://jl.ly ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: Last Call: draft-melnikov-smtp-priority-07.txt (Simple Mail Transfer Protocol extension for Message Transfer Priorities) to Proposed Standard
The most significant item that needs to be called out is the issue of tunneling the PRIORITY value through non-conforming MTAs by turning it into a message header field (MT-Priority) and then back again. This is a problematic technique, but is an important capability for those who need and intend to implement this extension. It creates a trust issue, wherein a message containing MT-Priority can be originated with a Message Submission Agent that does not know about this extension, and when the message hits a Message Transfer Agent that does support this, the header field will be turned back into a valid PRIORITY value, on the unwarranted assumption that it was authorized. Intermediate MTAs have no way to distinguish this situation from one where the field was tunneled legitimately. There may not be substantial experience with doing this sort of thing in SMTP relays, but there's plenty of experience doing it in gateways to other mail environments, e.g., X.400 and many of the old LAN email systems. In fact one of the more common fields that has been mapped this way is message transfer priority, so there is considerably experience with fields having more or less the same semantics as what is being proposed here. I am unaware of any cases where this was abused, probably because increased transfer priority doesn't buy you all that much in most cases. Related but more user-visible features, e.g., importance (in the X.400 sense) have been known to be abused, however. That said, I think an important omission in this document is that it only allows MSA's to change message priorities to conform to site policy. MTAs should also be allowed to do this. Another issue is the silly prohibition against using Priority: and other header fields to set priority levels. What if is existing site policy is in fact to use those fields to determine message priority? The counter-argument is that the use case for this specification involves out-of-band trust relationships, and that such situations will be known and dealt with. I believe that limits the usability of those features on the open Internet, with other use cases. Not if MTAs aren't also allowed to enforce site policy. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: DNS RRTYPEs, the difficulty with
In the previous message it was suggested that people who use a hosted DNS service should switch if their service doesn't support Type SPF. The problem is that, as far as I'm aware, none of them do. Perhaps in some cases it's because the person or persons who want to publish SPF records are not the ones who make the decision about which hosted DNS service provider to use, and are unable to make a sufficient case for a change to those who do have the authority to make it. It's also the case that even when the ability to publish such records exists, the person or persons who want to do so may not have the necessary access or authority and have no way to get it. Only today we were dealing with a customer who was asking about what options our product provides for overriding entries in his own DNS domain with local information. Yes, you read that right: The companies DNS has bogus stuff in it that cannot or will not be corrected. Such requests are unfortunately not uncommon. All that said, a suggestion: Patrik published a pretty good list of most of the issues that arise when attempting to deploy new RRTYPEs. Some of those, like the lack of GUI pulldowns in some hosting provider web interfaces, are clearly beyond the IETF's purview to address. But others, like the formatting issues in zone master files that arise when a new RRTYPE is defined, are things we can address. (And no, external scripts to do the format conversion are not a sufficient solution to this problem.) So might I suggest that instead of arging about the operational realities of deploying new RRTYPEs - which appears to be turning into yet another retelling of the parable about the blind men and the elephant - we instead turn our attention to activities that would actually address parts of the problem. In particular, John Levine has already written a draft that attempts to address the formatting issues by defining a simple format extension language. That might be a good place to start. (I've already sent him my review comments in private email.) Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: Issues with prefer IPv6 [Re: Variable length internet addresses in TCP/IP: history]
Yes, the issues with an unconditional prefer IPv6 approach have been noted, and operating systems of the vintages you mention certainly deserved criticism. In fact this has been a major focus of IPv6 operational discussions, and lies behind things like the DNS whitelisting method, the happy-eyeballs work, and my own RFC 6343. Old news; unfortunately it means you need new o/s versions. Disabling 6to4 and Teredo unless they are known to be working well is a good start, however. Old news perhaps, but an unavoidable consequence of this is that the oft-repeated assertions that various systems have been IPv6 ready for over 10 years don't involve a useful definition of the term ready. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: Issues with prefer IPv6 [Re: Variable length internet addresses in TCP/IP: history]
On 02/23/2012 13:51, ned+i...@mauve.mrochek.com wrote: Old news perhaps, but an unavoidable consequence of this is that the oft-repeated assertions that various systems have been IPv6 ready for over 10 years don't involve a useful definition of the term ready. The OP specified IPv4 only network. I suspect that if he had IPv6 connectivity his experience would have been quite different. I happily use Windows XP on a dual-stack network, for example. And systems running these old OS versions never under any circumstances move from one network to another where connectivity conditions change. Riiight. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: Issues with prefer IPv6 [Re: Variable length internet addresses in TCP/IP: history]
On 02/23/2012 14:48, Ned Freed wrote: On 02/23/2012 13:51, ned+i...@mauve.mrochek.com wrote: Old news perhaps, but an unavoidable consequence of this is that the oft-repeated assertions that various systems have been IPv6 ready for over 10 years don't involve a useful definition of the term ready. The OP specified IPv4 only network. I suspect that if he had IPv6 connectivity his experience would have been quite different. I happily use Windows XP on a dual-stack network, for example. And systems running these old OS versions never under any circumstances move from one network to another where connectivity conditions change. Riiight. Brian already covered unconditional prefer-IPv6 was a painful lesson learned, and I'm not saying that those older systems did it right. What I am saying is that for most values of IPv6 Ready which included putting the system on an actual IPv6 network, they worked as advertised. Which brings us right back to my original point: This definition of ready is operationally meaningless in many cases. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: Issues with prefer IPv6 [Re: Variable length internet addresses in TCP/IP: history]
In message 01occ10b11tc00z...@mauve.mrochek.com, ned+i...@mauve.mrochek.com w rites: On 02/23/2012 14:48, Ned Freed wrote: On 02/23/2012 13:51, ned+i...@mauve.mrochek.com wrote: Old news perhaps, but an unavoidable consequence of this is that the oft-repeated assertions that various systems have been IPv6 ready for over 10 years don't involve a useful definition of the term ready. The OP specified IPv4 only network. I suspect that if he had IPv6 connectivity his experience would have been quite different. I happily use Windows XP on a dual-stack network, for example. And systems running these old OS versions never under any circumstances m ove from one network to another where connectivity conditions change. Riiight . Brian already covered unconditional prefer-IPv6 was a painful lesson learned, and I'm not saying that those older systems did it right. What I am saying is that for most values of IPv6 Ready which included putting the system on an actual IPv6 network, they worked as advertised. Which brings us right back to my original point: This definition of ready i s operationally meaningless in many cases. I contend that OS are IPv6 ready to exactly the same extent as they are IPv4 ready. This isn't a IPv6 readiness issue. It is a *application* multi-homing readiness issue. The applications do not handle unreachable addresses, irrespective of their type, well. The address selection rules just made this blinding obvious when you are on a badly configured network. That's because the choice has recently been made that the place to deal with this problem is in applications themselves. I happen to think this is an exceptionally poor choice - the right way to do it is to provide a proper network API that hides network selection details, rather than demanding every application out there solve the problem separately. And yes, I'm familiar with the line of reasoning that says applications are too varied in their needs, or their internal environments conflict with the necessary use of threads, or whatever. I don't buy any of it. Yes, such applications exist, but like most things there's a sweet spot that solves a significant fraction of the problem. No one expect a disconnected IPv4 network to work well when the applications are getting unreachable addresses. Why do they expect a IPv6 network to work well under those conditions? First, you're comparing apples and oranges. Losing all network connectivity is very different thing from losing partial network connectivity. Nobody expects an applications that's totally dependent on the network to work without connectivity. That's just silly. But with other sorts of applications, I *do* have that expectation. I often work from places without any network connectivity using applications that employ networking as part of their function but also do non-networked stuff, and pretty much without exception they handle the transition fine, and have done so for years. Then again, I use a Mac. I have no idea what the state of play of Windows or Linux apps is in this regard. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: Last Call: draft-weil-shared-transition-space-request-14.txt (IANA Reserved IPv4 Prefix for Shared Address Space) to BCP
I support this updated draft, and I am keen for this to be published as a BCP. +1 I believe the amendments in this revision clarify the usage and intended purpose of the shared transition space. +1 Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: Last Call: draft-weil-shared-transition-space-request-14.txt (IANA Reserved IPv4 Prefix for Shared Address Space) to BCP
To the addressed folks who's messages appear below: I'm not sure I understand what you're saying. There was some objection at the beginning of this thread by Wes George, Noel Chiappa, and Brian Carpenter. I agreed that the document could be misunderstood as encouraging the use of the space as 1918 space and proposed some replacement text. There seemed to be some agreement around that text. Are you now objecting to that replacement text and want -14 published as is? Do you think the document should say that the new allocation can be used as 1918 space? If so, please explain. Not sure how a +1 to a statement saying I support this updated draft, and I am keen for this to be published as a BCP. can be interpreted in any but one way, or for that matter how it can be stated much differently. Anyway, to use different words, I would like to see the current draft approved and published as a BCP. Clear enough? Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: Last Call: draft-weil-shared-transition-space-request-14.txt (IANA Reserved IPv4 Prefix for Shared Address Space) to BCP
On 2/14/12 1:50 PM, ned+i...@mauve.mrochek.com wrote: Are you now objecting to that replacement text and want -14 published as is? Do you think the document should say that the new allocation can be used as 1918 space? If so, please explain. Not sure how a +1 to a statement saying I support this updated draft, and I am keen for this to be published as a BCP. can be interpreted in any but one way, or for that matter how it can be stated much differently. Anyway, to use different words, I would like to see the current draft approved and published as a BCP. Clear enough? Nope. Perhaps my question was unclear. I'll try and ask my question again with different words: Do you, or do you not, object to the proposed change that changes the text from saying, This space may be used just as 1918 space to This space has limitations and cannot be used as 1918 space? Nobody on the list objected to that new text. That new text significantly changes -14. You have stated your support for -14. Do you object to changing the text? I assumed the proposed change was in the current draft - isn't that how it's supposed to be done? Anyway, I think the change is fine, but I also think the draft is acceptable without it. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: the success of MIME types was Re: discouraged by .docx was Re: Plagued by PPTX again
- Original Message - From: Julian Reschke julian.resc...@gmx.de To: Yaakov Stein yaako...@rad.com Cc: John C Klensin john-i...@jck.com; ietf ietf@ietf.org Sent: Monday, November 28, 2011 6:07 PM On 2011-11-26 21:52, Yaakov Stein wrote: That leaves ASCII, a few forms of PDF, and RFC 5198-conforming UTF-8. That wouldn't bother me much, but be careful what you wish form. What we have been told is that the rationale behind the use of ASCII and several other formats is that they will remain readable on devices that will be used X years hence. ASCII is already unreadable on many popular devices and in a few years will be no better than old versions of word. ... Can we *please* distinguish between the character encoding we use (US-ASCII) and the file format (text/plain)? If *we* don't get this right, how can we expect anybody else to get it right? You will be aware of the recent threads on apps-discuss about MIME types (of which the text/plain you mention is one) which concluded, AFAICS, that there is no rationale why a (top level) type should or should not exist, there are no criteria for creating new ones, that it is impossible to draw up a taxonomy of types because there is no underlying logic in any dimension. And I would have to characterize all of that as 100% grade A hogwash. I'm not going to bother refuting each point since even if this nonsense were true top-level type rules, or the lack thereof are completely irreleant to the matter at hand. But feel free to read the thread if you want the real story. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: An Antitrust Policy for the IETF - why?
+1 to all of John's points here. Especially about the essential nature of lawyers - I've worked with plenty of them as well. Ned The IETF legal counsel and insurance agent suggest that the IETF ought to have an antitrust policy. I would be interested in a brief explanation of why we need one now, since we have gotten along without one for multiple decades. Having worked with a lot of lawyers, my experience is that few lawyers understand cost-benefit tradeoffs, and often recommend spending unreasonably large amounts of money to defend against very remote risks. Similarly, insurance agents will usually tell you to insure against anything. (This is why NDAs are 12 pages long, and the standard deductible on policies is usually an order of magnitude too small.) I don't know the particular lawyer and agent involved, and it's possible they're exceptions to the rule, but before spending much money, I would want to understand better what problem we are trying to solve and what the realistic risk is. Also keep in mind that the main effect of such a policy would be to shift whatever the risk is from the IETF onto participants. It might also be educational, here's things that might lead to personal legal risk if you talk about them, but we don't need a formal policy for that. I understand that some other SDOs have antitrust policies, but they generally have organizational members, and other differences from the IETF that make them only weakly analogous. R's, John ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: The death John McCarthy
First, as someone who chartered the working group, who has implemented Lisp (the programming language) at least four times, and who views Dr. McCarthy as a hero I disagree that name is problematic or disrespectful. And I almost take offense in the claim that this is a generational thing. I didn't charter the group and I've only done two Lisp implementations, but aside from that, +1. (Or should I say t instead?) And frankly, if there's disrespect to be found here, IMO it lies in using this sad event as a proxy to criticize some IETF work some people apparently don't like. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: Gen-ART review of draft-ietf-eai-rfc5335bis-12
On 10/20/2011 3:37 PM, Pete Resnick wrote: So, if the limit is still 998, then there is no change with respect the former limit. See the next sentence: (Note that in ASCII octets and characters are effectively the same but this is not true in UTF-8.) Remember, in UTF-8, characters can be multiple octets. So 998 UTF-8 encoded *characters* are likely to be many more than 998 octets long. So the change is to say that the limit is in octets, not in characters. The switch in vocabulary is clearly subtle for readers. (I missed it too.) I suggest adding some language that highlights the point, possibly the same language as you just used to explain it. The document already contains text for this: Section 2.1.1 of [RFC 5322 limits lines to 998 characters and recommends that the lines be restricted to only 78 characters. This specification changes the former limit to 998 octets. (Note that in ASCII octets and characters are effectively the same but this is not true in UTF-8.) The 78 character limit remains defined in terms of characters, not octets, since it is intended to address display width issues, not line length issues. Note the parenthetical comment in the middle of the paragraph. I really don't see any point in further elaboration of this issue in this context. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: how to make Apple Mail comply to email standards: or vv?
I've just received an important email from an important person (a former IETF chair even) on an IETF mailing list. The content-type was: Content-Type: text/plain; charset=us-ascii yet the first line of the mail was 491 characters wide. ... which is legal according to the relevant standards. There are legitimate uses for fixed format messages with long lines, which is why the 78 character wrap limit is a SHOULD, not a MUST. Of course in the case you present the line should have been wraped. But that's a user agent quality issue, albeit a fairly common one. There is an expectation that format=flowed; is now implied. That particular wrongminded expectation has been with us in one form or another for almost 20 years - it became a real problem shortly after MIME was published, as a matter of fact. There have been many attempts to address it, including the relatively recent definition of the format=flowed label. Has the YAM WG taken this into consideration? Of course it hasn't. Read the charter. YAM has a single purpose: The Yet Another Mail (YAM) WG will revise existing Internet Mail specifications currently at Draft Standard to advance them to Full Standard. Even if there was a consensus to make format=flowed the default (and I've seen no sign of any such thing), YAM couldn't do it. You know, that whole pesky no significant changes when moving from draft to full thing? Will a new standard now make make format=flowed; standard? Seems unlikely. I'm really really really tired of this. I never expected redmond to do the right thing, but at least everyone with a clue knew that they did the wrong thing, and some versions of that system there was even settings that mostly did the right thing. Meanwhile a colleague of mine in the non-IT space is complaining that he can't read email that I send because his Apple Mail program seems to forceably reformat all text, assuming it is format=flowed; RFC2646 has been around for 11 years now. So my request is simple: 1) can the IETF secretariat please document appropriate settings for commodity mail user agents that comply with RFC2464, and configure our mailing list software to filter anything that does not comply? I'm serious about this. It's embarassing. The first part seems reasonable, the second seems like a massive overreaction and IMNSHO has zero chance of being adopted. 2) alternatively, will the YAM (or another WG) please update the standard to say that format=flowed is now the default, and ideally also give me some idea if format=fixed will do anything at all with MUAs out there? Again, making such a major change to the default handling of the most comon type of email seems unlikely to be feasible at this stage. And besides, we've tried standards-making in this space quite a few times (simple text proposal, text/enriched, format=flowed, etc.) and nothing has worked. I see no reason to believe that further standards-making will do anything but further fragment the behavior of user agents. So perhaps it's time to try a different approach. Instead of writing yet another technical specification, write a best practices document that explains the problem and gives specific advice to user agent developers. This sort of thing has been conspicuous by its absence in previous documents. Just a thought. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: Last Call: draft-melnikov-mmhs-header-fields-04.txt (Registration of Military Message Handling System (MMHS) header fields for use in Internet Mail) to Informational RFC
On 2011-09-15 18:46, Mykyta Yevstifeyev wrote: ... 9) I'd like you hereby disallowed further registration of header fields beginning with MMHS, likewise RFC 5504 Downgraded prefix (http://tools.ietf.org/html/rfc5504#page-18). ... No. It's a bad idea in RFC 5504, and it would be a bad idea here. Agreed. Bad idea. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: Conclusion of the last call on draft-housley-two-maturity-levels
to actually perform such analysis without actually trying things at some sort of scale. I'm sorry, but I've seen no evidence that the necessary skills for this actually exists. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: Conclusion of the last call on draft-housley-two-maturity-levels
I find it impossible to believe that this will not result in even more hard-line positions on the part of some IESG members when something with which they disagree is a candidate for PS. I see no way in which the draft solves this problem, which remains one of its implicit goals. I said before, I don't care if it is published, because I think it will have little effect. But I think we'd better be prepared for some IESG members to insist on the same high bar for PS that we have under RFC 2026, regardless of what the RFC says. +1 Best statement of the problem with this document that I've seen so far. Except for one small problem: It's nonsensical. Why is it nonsensical? Because you're comparing the status quo with a possible future course of action. The one thing that's we can be certain of is things won't remain the same. They never do. So in order to make a reasonable comparison you have to project what's likely to happen if this document isn't approved, then compare that with what might happen if it is. And the future where this isn't approved is fairly easy to predict: As more and more documents become proposed standards and then fail to progress along the standards track - and the trend lines for this could not be clearer - we effectively move more and more to a one-step process. The IESG has only one rational response to that: Continue to raise the bar on the initial step to proposed. Will the imposition of a two step process change this? It certainly won't do so immediately, so the likely outcome is that yes indeed, the bar will continue to go up, at least initially, irrespective of whether or not this document is approved. But if more documents start advancing - and yes, that's an if - that will lessen the pressure on the initial step, and perhaps break the cycle we're currently stuck in. And please don't try trotting out a bunch of additional what ifs about how if this proposal fails we can then get past this distraction (or however you would characterize it) and address whatever it is you think the actual problems are. Given the time that has gone into trying to make this one simple and fairly obvious change, if it fails, the odds of someone attempting even more fundamental changes are pretty darned low. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
RE: Conclusion of the last call on draft-housley-two-maturity-levels
In looking through this discussion, I see: - People saying that moving from 3 steps to 2 steps is a small step in the right direction, lets do it. Many people who have said this (including I) have been silent for a while quite possibly because they have gotten frustrated with the endless discussion. Ross, I'm right there with you. I fully support this document at worse a small incremental step that clears away the some brush (at best it may actually turn out to be quite valuable) and I'm completely frustrated that this discussion is continuing. This really needs to stop now. And yes, some people aren't happy with the outcome. Thems the breaks. - People saying that there are other more important problems that we should be focusing on. Therefore, rather than either making this simple change or discussing other possible improvements in the process, instead let's debate this simple step forever and never get anything done. At least part of the problem is lack of agreement on what the issues are. Even if this step is a waste of time - I think that's unlikely but let's suppose - at least it will make it clear where the problems *aren't*. - People saying that this step won't do anything. Two things that I don't seem to have picked up on: (i) Any consensus that a 3 step process is better than a 2 step process; (ii) Any hint of moving towards an agreement on other things that we might do to improve the process. Well, that's the real problem, isn't it? Even if you believe this is a distraction and even actively harmful, it's not like we've been able to move past it either. The running code result here seems pretty clear, and it does not argue in favor of another round of discussion. I think that we should go to a two maturity level process, be done with this little step, and also enthusiastically encourage people to write drafts that propose *other* changes to the process. Then at least we can be debating something different 6 months from now than we were debating last year. +100 Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: Conclusion of the last call on draft-housley-two-maturity-levels
On Sep 2, 2011, at 5:36 PM, ned+i...@mauve.mrochek.com wrote: In looking through this discussion, I see: - People saying that moving from 3 steps to 2 steps is a small step in the right direction, lets do it. Many people who have said this (including I) have been silent for a while quite possibly because they have gotten frustrated with the endless discussion. Ross, I'm right there with you. I fully support this document at worse a small incremental step that clears away the some brush (at best it may actually turn out to be quite valuable) and I'm completely frustrated that this discussion is continuing. This really needs to stop now. And yes, some people aren't happy with the outcome. Thems the breaks. As far as our process is concerned, the question is, do we have rough consensus to accept it? I think it's dubious that we have such consensus, and apparently so do others. Simply put, I've watched the responses to this fairly closely, and I completely disagree with your assessment. Personally I think this proposal is Mostly Harmless, so I'm willing to hold my nose about it. But I'm very concerned about the argument that the default assumption should be that we change our process even in the absence of consensus to do so. Regarding the proposal, I get the impression that people are mostly in three camps: Well, none of these describe my own position, which is that eliminating the three step process will at a minimum act as an incentive to move more documents along. (You, and most others engaging in this debate, routinely neglect the psychological factors involved.) I can easily name a dozen RFCs, all currently at proposed, that I for one will be strongly incented to work to advance if this step is taken. And there isn't a chance in hell that I'll bother with any of them if this step doesn't happen, especially after the recent debacle surrounding the attempt to move 4409bis to full standard, and more generally given how the entire YAM experiment played out. I'm sorry, but passing down the advancement gauntlet is plenty hard enough to do once. Requiring it be done twice? Been there, done that, not remotely interested in doing it again. Additionally, by simplifying the process, we will gain essential insight into where other problems lie. Without such simplification I see no chance at all at making progress on any of these issues. 1) Even if this is a baby step, it's a step in the right direction. Or even if it's not a step in the right direction, taking some step will at least make it possible to make some changes in our process. Maybe we'll not like the results of taking this step, but at least then we'll have learned something, and if the result is clearly worse we'll be motivated to change it. (I call this change for the sake of change) That last substantially and obviously mischaracterizes this position. In fact I strongly recommend that you stop trying to summarize complex position with cute - and utterly wrong - phrases like this. This is annoying and quite unhelpful. 2) Fixing the wrong problem doesn't do anything useful, and will/may serve as a distraction from doing anything useful. (I call this rearranging the deck chairs) 3) People should stop arguing about this and just hold their noses about it, because the arguing will make it harder to do anything else in this space. (I call this Oceana has always been at war with Eurasia. Ok, that's probably too harsh, but it's what immediately comes to mind.) Actually, I think there are a substantial numer of people who believe exactly the opposite of this. All of these are defensible theories.As it happens, I don't believe #1 applies in this space, I do believe #2, and I have to admit that #3 does happen. The arguments that people are giving in favor of approving this bother me more than the proposal itself does. (I'm a firm believer that good decisions are highly unlikely to result from flawed assumptions, and flawed assumptions often affect many decisions. So challenging a widely-held flawed assumption is often more important than challenging any single decision.) Well, the main argument I'm giving is based on my own perception of the effect this will have on myself and similarly minded people as a contributor. If you think that assessment is incorrect, then I'm sorry, but I think you're being extraordinarily foolish. ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: Conclusion of the last call on draft-housley-two-maturity-levels
First, I'm in full agreement with Ross. Second, for the record and as a response to Keith, my read of the discussion on the last call was the biggest group of responses said that we should move forward with the draft. There were two smaller groups, those with a clear objection and those with roughly a no-objection or it does not cause harm opinion (and a group who seemed to discuss orthogonal issues and not respond to the question). I could of course have made mistakes in this determination, but I thought it was rough (perhaps very rough) consensus. FWIW, this matches my own assessment almost exactly. Of course, it gets more interesting if you start thinking about the reasons why people wanted to move forward. Keith's latest e-mail has interesting theories about those. I don't think anyone thinks this is the priority #1 process fix for the IETF. Agreed. For me, cleaning cruft from the IETF process RFCs is a big reason for supporting this work. And I must admit that we seem to be in a place where its very, very hard to make _any_ process RFC changes. Getting one done, even if its a small change would by itself be useful, IMO. Finally, I think two levels are enough. Cruft elimination is also a Good Thing. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: Conclusion of the last call on draft-housley-two-maturity-levels
no chance at all at making progress on any of these issues. Okay, I can see that as a possibility. Sometimes when undertaking a great task, it doesn't matter what subtask you pick to do next, as long as you do something. Momentum is often more important than doing things in order of importance. My question is then, how many people think that we need to undertake a great task where our process is concerned, and how many of those think that given current political conditions, if we undertake such a task, we're likely to end up with something substantially better than we have now? (I'm open to the idea but skeptical) My answer is I have no idea. Our current process is so complex - and unnecessarily so - that every time we even try and discuss substantial changes the discussion goes off in about fifty different directions. When it comes to the really big stuff we can't even agree on where to start, let alone get consensus on what to try and fix. So let's please take the small step of simplifying the process a little first, so perhaps we can get some perspective on the big stuff. Or not - it may well be that this small step isn't sufficient to gain any sort of perspective, but I've already given my reasons for believing it's useful in it's own right even if that does not happen. 1) Even if this is a baby step, it's a step in the right direction. Or even if it's not a step in the right direction, taking some step will at least make it possible to make some changes in our process. Maybe we'll not like the results of taking this step, but at least then we'll have learned something, and if the result is clearly worse we'll be motivated to change it. (I call this change for the sake of change) That last substantially and obviously mischaracterizes this position. In fact I strongly recommend that you stop trying to summarize complex position with cute - and utterly wrong - phrases like this. This is annoying and quite unhelpful. There are definitely cases where a journey of a thousand miles begins with a single step, I'm just skeptical that that argument applies in this specific case. 2) Fixing the wrong problem doesn't do anything useful, and will/may serve as a distraction from doing anything useful. (I call this rearranging the deck chairs) 3) People should stop arguing about this and just hold their noses about it, because the arguing will make it harder to do anything else in this space. (I call this Oceana has always been at war with Eurasia. Ok, that's probably too harsh, but it's what immediately comes to mind.) Actually, I think there are a substantial numer of people who believe exactly the opposite of this. I'm not sure I understand what you mean here. Are you saying that there are a substantial number of people who wish to make it harder to do anything at all in this space, so they keep arguing about it? Or something else? I think a substantial number of people believe there is nonegligable benefit in doing this. No nose-holding required. The arguments that people are giving in favor of approving this bother me more than the proposal itself does. (I'm a firm believer that good decisions are highly unlikely to result from flawed assumptions, and flawed assumptions often affect many decisions. So challenging a widely-held flawed assumption is often more important than challenging any single decision.) Well, the main argument I'm giving is based on my own perception of the effect this will have on myself and similarly minded people as a contributor. If you think that assessment is incorrect, then I'm sorry, but I think you're being extraordinarily foolish. I think you're in an excellent position to understand how approval or disapproval of this document will affect your interest in doing work on the documents you mentioned, and I'm sure you're not the only one who would be encouraged by such a change to our process. And IMO that's a good thing. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
RE: Minimum Implementation Requirements (Was: 2119bis)
-Original Message- From: ietf-boun...@ietf.org [mailto:ietf-boun...@ietf.org] On Behalf Of Melinda Shore Sent: Thursday, September 01, 2011 12:45 PM To: ietf@ietf.org Subject: Re: Minimum Implementation Requirements (Was: 2119bis) Can anybody point to an incident in which lack of clarity around 2119 language caused problems, and it was determined that 2119 itself was the problem and not authors or editors being careless? +1. As we've defined SHOULD and MUST in RFC2119, they lay out conformance requirements. I still don't see what's broken. If the Why is this a SHOULD and not a MUST? question that Spencer pointed out is a common one, then guidance to authors might be an appropriate addition. The fact that reviewers sometimes question the choice between a SHOULD and MUST (in either direction) demonstrates absolutely nothing in and of itself. In fact I see such questions as entirely healthy and appropriate, indeed, if they didn't come up fairly regularly I'd strongly suspect we have a much bigger problem with our specifications not taking operational realities into account. Now, if when such questions arise the eventual outcome is often that the SHOULD gets changed to a MUST or vice versa, then we may have an issue with using the terms that needs to be addressed. Or we may not - there are plenty of other possible explanations. But we can save the root cause analysis until there is at least some evidence that compliance terms are regularly being *changed* as a result of review feedback. AFAIK such evidence has not been introduced. Ned P.S. And if anyone seeks to provide such evidence, please remember that data is not the plural of anecdote. ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
re: Pachyderm
can you please explain *why* publishing conformance statements would be such a bad idea? I am not being cynical, I really want to understand the reasoning. (I don't know Pete's reasons, but I suspect they're not dissimilar from my own. Which are ...) The main problem with conformance languge is that conformance has a nasty way of becoming an end unto itself, and the *reasons* why conformance is desired get lost along the way. The result is technical compliance to a bunch of words on paper that don't provide an actual, useful, result like, say, insisting on interoperability does. For example, the X.400 email standards are all about conformance. Incredibly elaborate and picky conformance test suites can be, and have been, written for this stuff. So how is it that, after passing a truly massive test suite that checked every last little conformance detail in the specifications (and paying lots of $$$ for the privilege), our software then failed to interoperate in at least half a dozen different ways with another piece of software that as it happened had also just passed the exact same test suite? Heck, we couldn't even get the thing to properly negotiate a session to transfer messages. And once we got that working (in the process we ended up having to violate a couple of those test suite requirements) we were immediately lost in a thicket of differing interpretations of not just protocol fields but even the basic elements that comprise X.400 addresses. And this is supposed to be useful? As a moneymaker for software developers, maybe - you may rest assured the cost of all of this nonsense were passed on to our customers, many of whom were bound by various regulations and had no choice but to buy this crap - but as a way to get things to work, most assuredly not. And this trap is a lot easier to fall into than you might think. I've fallen into it myself - I once spent entirely too much time dithering about getting severe error handling right in a particular programming language implementation, completely losing sight of the fact that once an error this severe occurs the options are limited and the outcomes are so poor it just isn't that important that the rules are followed to the letter. It made a lot more sense on getting rid of the conditions that could cause an error. And, for extra credit, what do you make of http://tools.ietf.org/html/rfc5996#section-4 (in my own backyard)? Well, the section may be titled Conformance Requirements but the section is all about interoperability, not conformance. So that's fine in my book. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: 2119bis
On Aug 30, 2011, at 9:24 AM, Marshall Eubanks wrote: I support adding the SHOULD ... UNLESS formalism (although maybe it should be MUST... UNLESS). It would be useful as there will be times where the UNLESS can be specified and has been given due consideration at the time of writing. That, however, will not always be the case. (More inline). How would SHOULD...UNLESS or MUST...UNLESS differ from using the current 2119 definitions and just writing SHOULD...unless or MUST ... unless? Personally I think 2119 is just fine and doesn't need to be updated. +1. I'm still not seeing sufficient justification to open this particular can of worms at this juncture. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: 2119bis
Hi - From: Eric Burger eburge...@standardstrack.com To: Narten Thomas nar...@us.ibm.com; Saint-Andre Peter stpe...@stpeter.im Cc: IETF discussion list ietf@ietf.org Sent: Monday, August 29, 2011 3:08 PM Subject: Re: 2119bis I would assume in the text of the document. This paragraph is simply an enumeration of Burger's Axiom: For every SHOULD, there must be an UNLESS, otherwise the SHOULD is a MAY. I disagree. I concur with your disagreement. SHOULD should *not* be used when the list of exceptions is known and practically enumerable. If the UNLESS cases can be fully enumerated, then SHOULD x UNLESS y is equivalent to WHEN NOT y MUST X. (Both beg the question of whether we would need to spell out that WHEN y MUST NOT X is not necessarily an appropriate inference.) RFC 2119 SHOULD is appropriate when the UNLESS cases are known (or suspected) to exist, but it is not practical to exhaustively identify them all. Let's not gild this lily. +1 Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: https
On 8/27/2011 4:12 PM, t.petch wrote: Glen Me again. Just after I posted my last message, I received a post on the ietf-ssh list, hosted by netbsd.org, and looking through the 'Received: from' headers, as one does on a wet Saturday morning of a Bank Holiday weekend, there was TLS, used to submit the message to the server, so even spammers have caught on that TLS should be used everywhere. End to end? As a side note, the reason spammers use TLS to submit mail is obvious: It's required by many submission servers so they really don't have a choice. (The reason it's required is to protect the authentication exchange, not because there's any real expectation that it provides useful privacy protection for the submitted email itself.) Apparently, from the POV of the spammer his SMTP server. Email is a store forward system. In any case, my original question was not about the definition of end-to-end, it was about Ned usage of the term hop. I used the term hop in a very generic sense to refer to moving the data around. Upon further analysis, however, it seems clear that he was referring to the email archives as if they are something other than simple files (as betrayed by his statement that Don't pretend a transfer protection mechanism covering exactly one hop provides real object security, because it doesn't.); thus, the retrieval of the archived data would be the last hop in the email system. And that's incorrect. For one thing, I often retrieve material from web sites and save it rather than looking at it right there on the screen. So the transfer of the material from the web server is in no way, shape, or form the final hop the information takes before it is consumed. As as Keith points out, I and many others am often forced to do such access through corporate-mandated proxies of various sorts - another hop. There seem to be two problems with this statement: one is taking the file transfer mechanism as if it was part of the email system itself, Nobody is making any such claim. which it obviously is not (downloading an archived message is no different than downloading an RFC from a Web site); the other being that someone, somewhere was pretending that TLS does something that it was never designed to do (a straw man of, AFAICT, Ned's invention -- I don't recall anybody making such a claim on this thread, I was responding to the justification given for the use of https in this context. The exact words used were: The mail archives (and the minutes of the physical meetings) are the official record of the Working Groups, IETF, etc. Those archives should be available with a reasonably high level of integrity and authenticity. Nor was I the only, or even the first, to suggest that object security is needed for this sort of protection. nor for that matter saying they _wanted_ real object security applied to the archives, merely that it's not really a bad idea for a person retrieving them to have some assurance that they came from the IETF server and that they weren't modified in transit). And once again, nobody is saying that TLS doesn't give some very limited assurance along these lines - the notion that there are claims to the contrary is your own strawman. What we are saying is that there are also significant costs, those costs appear to be greater than the benefits in this case, and if there is real concern about archive integrity there are better ways to secure them. Anyway, this discussion is now well past it's sell-by date, so this will be my final response on the topic. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: https
I'm increasingly coming to think that *everything* should be done with TLS unless you can prove it's harmful. Call me paranoid, but given the general state of the world, secure-by-default seems like the way to go. -Tim This sentiment always sounds nice, but the devil is in the details. Back when STARTTLS was added to SMTP, a bunch of us thought this provided the means to do opportunistic encyption of email transport, so there was a push to deploy that. It was a total failure - it seems a certain large vendor with a very popular implementation borked their server so it advertised STARTTLS even though no certificate was installed and any attempt to negotiate TLS protection would end in failure. And since the negotiation failure left the connection in an unusable state, a new connection had to be tried, along with all the logic to prevent the new one from ending up in the same state. It tooks years before enough of the broken servers were fixed to try again, and by then service defaults and deployment guidelines were well established and the opportunity was gone as well. Of course this set of issues was unique to the situation. But with TLS it's always something - if it isn't broken default configurations, its incompatible cipher suites, or problems with certificate formats, or expired certificates, or dubious CAs, or major security holes in popular implementations, etc. etc. The bottom line is this stuff is just too complicated for mere mortals to get right, and the fact that they then proceed to get it wrong more often than not causes an enormous amount of trouble. In the present case, for example, you may think that an expired certificate is not a big deal. Not so. For one thing, the more people have to click on the this really isn't secure, proceed anyway button to get things done, the more likely they are to do it when it's a real attack and not a maintenance problem. And for another, there are some browser/certificate error combinations where there's no workaround - no amount of clicking I agree buttons will get you the data you're after. I don't have an anwwer here, but the one thing I'm fairly sure of is that blindly pushing TLS everywhere is not the solution a lot of folks believe it is. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: https
On 8/27/11 7:25 AM, ned+i...@mauve.mrochek.com wrote: I don't have an anwwer here, but the one thing I'm fairly sure of is that blindly pushing TLS everywhere is not the solution a lot of folks believe it is. I tend to think that the problem here (and I agree that it's a big one) isn't TLS, but that PKI as defined by pkix is very difficult to deploy correctly. Agreed. I've seen similar sorts of problems with digital signatures on email, but in those cases as often as not someone simply got the certificate contents wrong (or the user doesn't understand how to configure his mail client correctly and is using a name that doesn't appear in the certificate) rather that the cert has expired (although there's a lot of that, too). There's a substantial usability problem. Absolutely, and it's both architectural and operational - PKI is full of complex and subtle concepts that implementations don't exactly help you with. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: https
--On Friday, August 26, 2011 09:43 -0400 Donald Eastlake d3e...@gmail.com wrote: Yup, but why are we using https at all? Who decided, and please would they undecide? Unexpired certificates can be circumvented, but all too often, the https parts of the web site just do not work and, more importantly, I think it wrong to use industrial grade security where none is called for. The mail archives (and the minutes of the physical meetings) are the official record of the Working Groups, IETF, etc. Those archives should be available with a reasonably high level of integrity and authenticity. Don, If that is the goal, wouldn't we be lots better off just digitally signing those things, just as we are gradually starting to create signatures for I-Ds, etc.? Verifying that one is talking to the right server and that the content is not tampered with in transit is all well and good, but it doesn't protect against compromised documents or a compromised server at all. +1. If you want signatures, do them properly. Don't pretend a transfer protection mechanism covering exactly one hop provides real object security, because it doesn't. And as for the encrypt so the really secret stuff doesn't stand out argument, that's fine as long as it doesn't cause inconvenience to anyone. That's clearly not the case here. And I'm sorry, the mistakes were made notion doesn't really fly: Certificates aren't a set it and forget it thing, so if you haven't noted expiration dates on someone's to-do list so they can be updated before expiration, you're not doing it right. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: https
On Fri, Aug 26, 2011 at 11:14 AM, ned+i...@mauve.mrochek.com wrote: +1. If you want signatures, do them properly. Don't pretend a transfer protection mechanism covering exactly one hop provides real object security, because it doesn't. It ensures that what you're getting is the same as what the IETF has on file, No, it really doesn't do that. There can be, and usually are, a lot of steps involves besides the one https hop. and (assuming you trust the IETF's archive integrity, which is a separate problem) On the contrary, it's the same problem. You're just pretending that solving an insignificant subset of that problem is useful. It isn't. what everyone else on the list received. This, OTOH, *is* a separate problem, one that isn't addressed in any way by https on archive archive access. Nor does a signed archive address this issue. Nonrepudiation of delivery is in fact a very difficult problem to solve - the algorithms I'm aware of for it either involve a TTP or are exceptionaly ugly - and the fact that the architecture of our email services isn't designed for it doesn't exactly help. (In case you care, it's the nontransactional nature of POP and IMAP that gets in the way of doing nonrepudiation of delivery at the email protocol level.) We're very lucky that we don't operate in a regime where this is actually a requirement. (And such regimes do exist, although the solutions they actually use tend to be hacks.) It seems to me it's more important to know that than to know that stuff sent to the list actually came from who it claims to come from. Does it really matter if a proposed standard wasn't really designed by who the archive says it was designed by? And now you're talking about yet a third problem - nonrepudiation of submission. It's completely doable, but only by significantly increasing the barriers to participation by requiring signed messages. I don't believe the benefits come even close to the costs. And as for the encrypt so the really secret stuff doesn't stand out argument, that's fine as long as it doesn't cause inconvenience to anyone. That's clearly not the case here. And I'm sorry, the mistakes were made notion doesn't really fly: Certificates aren't a set it and forget it thing, so if you haven't noted expiration dates on someone's to-do list so they can be updated before expiration, you're not doing it right. Yeah, it seems like the IETF would be on top of certificate expiration, having invented it. But I think having the encryption is a very good thing. Some government interests would be happy to keep information about some aspects of IETF business away from their citizenry, and have the resources to launch a MITM attack. (Of course, those governments may also have the resources to sign a fake certificate, but that is, again, a separate problem.) I'm sorry, but the notion that the present use of https provides any sort of protection against this sort of thing is just silly. The simple facts that the archives are also available via regular http and that there is no stated policy about the availability of archives via https means the present setup is completely vulnerable to downgrade attacks. Again, if you really care about this sort of stuff, the archives need to be timestamped and signed. Once the certificate expiration business is fixed, it should be fairly simple to make sure it's kept up to date so that this sort of thing doesn't happen again. 10+ years of experience with multiple certificates having multiple failures says this is, at best, a crapshoot. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
Re: Last Call: draft-ietf-intarea-ipv6-required-01.txt (IPv6Support Required for all IP-capable nodes) to Proposed Standard
I find this document utterly bizarre and think it would seriously damage the Internet to publish it. This seems a little ... extreme. The document appears to me to be Mostly Harmless, with all that implies. The idea that ipv6 should be regarded as normal, as of equal standing to ipv4 is fine, the sort of statement that the IAB should make, or have made, as an RFC or in some other form. But this I-D claims Updates [RFC1122] to clarify that this document, especially in section 3, primarily discusses IPv4 where it uses the more generic term IP and is no longer a complete definition of IP or the Internet Protocol suite by itself. IPv4 is a phenomenal success, and RFC1122 is a key part of that. IPv4 was a confused jumble, as IPv6 is now, and RFC1122, with another two or so I-Ds, cut through the cruft and rendered it usable. IPv6 desparately needs an equivalent to RFC1122, Complete agreement on this point. Such a document, informed by actual IPv6 deployment experience at some sort of scale, is urgently needed. And this most certainly is NOT that document. But unless publishing this is seen as meeting the need for an 1122v6 - and I've seen no indication that's the case - I fail to see the harm. OTOH, if this really is seen as being a 1122v6, then I join you in opposing it's publication. as a trawl of the v6ops list archives shows, and clearly this I-D is never going to be it, but claiming that this I-D provides an update to RFC1122, coupled with its title, gives the message that there is not going to be such an I-D; IPv6 will remain a confused jumble (and so is unlikely ever to emulate the success of IPv4). Maybe I'm being clueless about this, but I don't see how IPv6 Support Required for all IP-capable nodes gives this impression. Ned ___ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf