Hi Paul,

On Wed, May 07, 2014 at 03:37:45PM -0700, Paul Vixie wrote:

> keep it civil, please.

I think I was being civil.  If you're going to quote simplistic
bromides from the pages of a comic book as the basis of your argument,
you should expect to be teased.  
 
> let me show you something wrong with that, then.
> 
> http://www.cisco.com/c/en/us/products/collateral/wireless/5700-series-wireless-lan-controllers/data_sheet_c78-722607.html
> 
> an rfc is seen by sellers and buyers as an unalloyed good. 

I'm not sure what you think is wrong with that page.  It is a product
page.  It tells you what it implements.  You are presumably supposed
to be an educated consumer who can make decisions about whether a
given feature is one you want.  Certainly, I do not believe everything
Cisco says, so I'm unlikely to believe them when they say, "Bees'
knees, man!  Just have some of this RFC!"

Is it the case that people often use, "It's an RFC," to be evidence
for, "It is good"?  Sure.  But first,
http://tools.ietf.org/html/rfc1796.  Second, people who are looking
for an excuse not to think do not have to work very hard at it.
Moreover, defending against the impulse you're worried about is one of
the reasons that nobody can ever get anything published around the
IETF, a problem that you have frequently complained about in public.

> the current thrust is to move from "ietf should do good engineering" to
> "ietf should document anything that has to be interoperable"

That is at once a straw person and a false dichotomy.  First, some of
us are arguing that part of good Internet engineering is in fact to
document stuff that we know people are deploying and that needs
interoperability.  That doesn't seem like such a controversial
proposition to me, so I'd be interested to hear why you think it to be
false.  Second, I don't think anyone is arguing that the IETF should
document _anything_ that has to be interoperable -- unless by "has to"
you mean something critical enough to the Internet (such as the DNS)
that, again, I'd have thought the proposition uncontroversial.

> The Domain Name System as originally contemplated and implemented made
> the assumption that a Full Service Resolver would either share a host
> with, or a LAN with, or a campus with its end-user stub resolvers.

I just had a quick look through 1034 and 1035, and I do not think that
the text offers very strong evidence for your claim there.  There is
admittedly the text in section 5.3.1 of 1034, "or can centralize the
cache for a whole local network or organization."  On the other hand,
there is the text in 5.1, "A very important goal of the resolver is to
eliminate network delay and name server load from most requests by
answering them from its cache of prior results.  It follows that
caches which are shared by multiple processes, users, machines, etc.,
are more efficient than non-shared caches."  This latter text
militates toward broadening the scope of a resolver.  Since many of
the "wide area full service resolvers" are in fact anycast and
therefore topologically close, and therefore able to eliminate
"network delay and name server load", it seems to me that the
architecture is entirely ambivalent on this topic.  So I'd appreciate
a citation.  Or, you know, a draft, "Wide Area Recursive Service
Considered Icky", or something.

> that's a terrible example since any internet draft describing a
> controversial technique either failed or got an applicability statement.

Really?  We had a _whole working group_ whose entire output was about
NAT and its tricks.  We have another one (MIF) that has suggested all
manner of hideous hacks in order to figure out which network you're
in.  And so on.  I dearly wish you were right that controversial
techniques always fail or get an AS.  

Anyway, I think I should be embarrassed to have a meta-argument about
whether we even ought to work on a draft.  Those of us who think
edns0-client-subnet ought to be documented can work on it, so that's
what I'm off to do.

Best regards,

A

-- 
Andrew Sullivan
[email protected]

_______________________________________________
DNSOP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dnsop

Reply via email to