Re: [whatwg] Absent rev?
Hello Ian, Ian Hickson wrote: On Tue, 18 Nov 2008, Martin McEvoy wrote: The second most common value was rev=stylesheet, which is meaningless and obviously meant to be rel=stylesheet. And that was the basis of the whatwg decision to drop rev? Yes. Was this the study you based your decisions on? http://code.google.com/webstats/2005-12/linkrels.html (I am not criticizing just trying to understand it) surely all it needed was to define some rev values (the same as rel) and people will start using rev correctly? That's backwards -- looking for a problem to fit the solution, not looking for a solution to fit the problem No not really because If you look at the anyalasis(link above) made in 2005 rev=made (9th) is used more than, rel start, search, help, top, up, author and a whole lot of other link relationships that have made their way into HTML5, It doesn't make any sense? If you have a more up to date study on link relationships, please can I have a link? Best Wishes -- Martin McEvoy http://weborganics.co.uk/
Re: [whatwg] Video element and duration attribute
2008/11/17 Ian Hickson [EMAIL PROTECTED]: It seems like a bad idea, since if we do this we'd have to define all kinds of error handling for when the author says the clip is 60 seconds long and then seeks to second 50 when the clip is really 40 seconds long, etc. Agreed. The likely use cases that could be perceived for this attribute are to allow the HTML author to: * specify a shorter duration to be played than what is actually available in the media * hint to the browser what the actual duration of the media is, for formats that don't say so up-front in media or HTTP headers * hint to the user how long the content is before they hit play and begin streaming (as displayed in the browser's normal timeline, without adding extra text to the surrounding page). I think these use cases are better handled with media fragments, which give a more precise hint about the subview being played. As for Ogg, we should (from xiph.org, not whatwg) encourage use of X-Content-Duration. I've make a note to add that to oggz-chop's response headers ... Conrad.
Re: [whatwg] Absent rev?
Martin McEvoy writes: Ian Hickson wrote: On Tue, 18 Nov 2008, Martin McEvoy wrote: (I am not criticizing just trying to understand it) surely all it needed was to define some rev values (the same as rel) and people will start using rev correctly? That's backwards -- looking for a problem to fit the solution, not looking for a solution to fit the problem No not really because If you look at the anyalasis(link above) made in 2005 rev=made (9th) is used more than, rel start, search, help, top, up, author and a whole lot of other link relationships that have made their way into HTML5, It doesn't make any sense? There's a difference between adding an attribute and adding to the set of values defined for an attribute; given rel's existence, the cost of adding start, up, etc is quite possibly less than of adding rev. There's also the misuse to consider. If, say, rel=up is barely used but when it is used it's generally used correctly then it's benign, and not causing any harm. Significant rev misuse has been identified; its existence is confusing people into writing something they don't mean. Smylers
Re: [whatwg] Workers and queue of events
On Wed, Nov 19, 2008 at 7:43 PM, Jonas Sicking [EMAIL PROTECTED] wrote: And if it becomes a problem we might in a future version be able to add something like a 'messagepostfailed' event that is fired on the sending port in case a message failed to reach its target for one reason or another. I don't think that would be useful. What if the message reaches the target but the target dies while the message is queued? Or it dies after executing the first JS statement in the message handler? The only way to be sure that a message has been processed to receive a response from the target saying so. If I was implementing workers in their own processes, I'd be tempted to make abnormal termination of the worker fatal to any Web page that was aware of the existence of the worker. The principle of if you can't follow the spec, destroy the evidence so no-one can prove it. Rob -- He was pierced for our transgressions, he was crushed for our iniquities; the punishment that brought us peace was upon him, and by his wounds we are healed. We all, like sheep, have gone astray, each of us has turned to his own way; and the LORD has laid on him the iniquity of us all. [Isaiah 53:5-6]
Re: [whatwg] Workers and queue of events
Nov 19, 2008, в 12:55 PM, Robert O'Callahan написал(а): And if it becomes a problem we might in a future version be able to add something like a 'messagepostfailed' event that is fired on the sending port in case a message failed to reach its target for one reason or another. I don't think that would be useful. What if the message reaches the target but the target dies while the message is queued? Or it dies after executing the first JS statement in the message handler? It turns out that message confirmations are necessary for garbage collection anyway - one can't GC a worker object if its thread/process has pending messages, or running scripts, because they can talk back and post events to worker.onmessage. So, we send a confirmation after the message has been dispatched, not just queued in the receiver. This is functionally equivalent to what Gecko does, I believe. Note however that I'm talking about worker objects here, not ports. - WBR, Alexey Proskuryakov
Re: [whatwg] Absent rev?
On Wed, 19 Nov 2008, Martin McEvoy wrote: Was this the study you based your decisions on? http://code.google.com/webstats/2005-12/linkrels.html That study was based on the first set of data I obtained, but I have since made many more detailed studies. (I am not criticizing just trying to understand it) surely all it needed was to define some rev values (the same as rel) and people will start using rev correctly? That's backwards -- looking for a problem to fit the solution, not looking for a solution to fit the problem No not really because If you look at the anyalasis(link above) made in 2005 rev=made (9th) is used more than, rel start, search, help, top, up, author and a whole lot of other link relationships that have made their way into HTML5, It doesn't make any sense? The problem solved by rev=made (or rel=author, which is the same) is the problem of how to indicate the author of the page. We have solved that problem in HTML5 (with rel=author). The idea of defining rev values because nobody uses rev is what I was referring to when I said that it was backwards. If you have a more up to date study on link relationships, please can I have a link? I have not published anything recently, but the results have not changed significantly since that 2005 study was published. On Wed, 19 Nov 2008, Martin McEvoy wrote: OK that makes sense, what cost is there of using rev and defining a few rev link types? Author confusion, implementation cost, testing cost, cost in writing tutorials, cost in writing validators, etc. This is the bit that I find so very wrong the most popular rev value is rev-made which is used correctly most of the time, Authors Misuse br all the time, the same goes for address based on the statement above HTML5 should drop those too? We are considering dropping address, though on balance it is used correctly quite a lot too, so it's not clear whether removing it would be better or worse overall. br we probably can't drop since it is used so widely and does have some pretty important valid uses. -- Ian Hickson U+1047E)\._.,--,'``.fL http://ln.hixie.ch/ U+263A/, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
Re: [whatwg] Absent rev?
Hello Philip Philip Taylor wrote: On Wed, Nov 19, 2008 at 9:35 AM, Martin McEvoy [EMAIL PROTECTED] wrote: [...] http://code.google.com/webstats/2005-12/linkrels.html [...] If you have a more up to date study on link relationships, please can I have a link? http://philip.html5.org/data/link-rel-rev.txt has some more recent data, from a different set of pages (and so with different biases, e.g. there's lots of Wikipedia and IMDB pages using rel=apple-touch-icon), with less processing (no case-insensitivity or token-splitting). Thank you Philip that is the most useful set of data I have seen for a long time It basically says that the whole premise that HTML5 should drop *rev* (a) because authors use it wrong, (b) Many authors use rev-stylesheet wrong, is a MYTH and an inaccurate assessment of the *rev* attribute Out of the 127249 pages studied, only 0.09% actually use rev=stylesheet Great stuff Thanks! -- Martin McEvoy http://weborganics.co.uk/
Re: [whatwg] Absent rev?
Smylers wrote: Martin McEvoy writes: Ian Hickson wrote: On Tue, 18 Nov 2008, Martin McEvoy wrote: (I am not criticizing just trying to understand it) surely all it needed was to define some rev values (the same as rel) and people will start using rev correctly? That's backwards -- looking for a problem to fit the solution, not looking for a solution to fit the problem No not really because If you look at the anyalasis(link above) made in 2005 rev=made (9th) is used more than, rel start, search, help, top, up, author and a whole lot of other link relationships that have made their way into HTML5, It doesn't make any sense? There's a difference between adding an attribute and adding to the set of values defined for an attribute; given rel's existence, the cost of adding start, up, etc is quite possibly less than of adding rev. OK that makes sense, what cost is there of using rev and defining a few rev link types? There's also the misuse to consider. If, say, rel=up is barely used but when it is used it's generally used correctly then it's benign, and not causing any harm. Significant rev misuse has been identified; its existence is confusing people into writing something they don't mean. This is the bit that I find so very wrong the most popular rev value is rev-made which is used correctly most of the time, Authors Misuse br all the time, the same goes for address based on the statement above HTML5 should drop those too? Smylers Thanks -- Martin McEvoy http://weborganics.co.uk/
Re: [whatwg] Absent rev?
On Wed, Nov 19, 2008 at 9:35 AM, Martin McEvoy [EMAIL PROTECTED] wrote: [...] http://code.google.com/webstats/2005-12/linkrels.html [...] If you have a more up to date study on link relationships, please can I have a link? http://philip.html5.org/data/link-rel-rev.txt has some more recent data, from a different set of pages (and so with different biases, e.g. there's lots of Wikipedia and IMDB pages using rel=apple-touch-icon), with less processing (no case-insensitivity or token-splitting). -- Philip Taylor [EMAIL PROTECTED]
Re: [whatwg] Absent rev?
Ian Hickson wrote: The problem solved by rev=made (or rel=author, which is the same) is the problem of how to indicate the author of the page. We have solved that problem in HTML5 (with rel=author). That does not solve the problem of rev=made because its not the same as rel=author author can relate to multiple instances on a page saying WE made this, an Author may have no control over who claims authorship of a page. made is usually a single point perspective, Its a way of authors claiming their own links in a statement saying I made This. rev=made is subtle but ever so important link relationship for an author, HTML5 really shouldn't abandon rev because of it. Thanks -- Martin McEvoy http://weborganics.co.uk/
Re: [whatwg] Absent rev?
On Wed, Nov 19, 2008 at 1:03 PM, Martin McEvoy [EMAIL PROTECTED] wrote: Philip Taylor wrote: http://philip.html5.org/data/link-rel-rev.txt has some more recent data, from a different set of pages (and so with different biases, e.g. there's lots of Wikipedia and IMDB pages using rel=apple-touch-icon), with less processing (no case-insensitivity or token-splitting). Thank you Philip that is the most useful set of data I have seen for a long time It basically says that the whole premise that HTML5 should drop *rev* (a) because authors use it wrong, (b) Many authors use rev-stylesheet wrong, is a MYTH and an inaccurate assessment of the *rev* attribute Out of the 127249 pages studied, only 0.09% actually use rev=stylesheet The premise from near the beginning of this thread was: We did some studies and found that the attribute was almost never used, and most of the time, when it was used, it was a typo where someone meant to write rel= but wrote rev=. I think that ought to say ... (excluding rev=made, which is uninteresting since it's redundant with rel=author) In that case, rev is used on 0.2% of pages, which justifies the claim almost never used. And rev=stylesheet makes up 57% of those uses of rev, which justifies the claim most of the time ... it was a typo (under a loose definition of typo that includes people copying-and-pasting without understanding the distinction between rel and rev, which is the impression I get from looking at some of these pages). And looking at some other values, e.g. link rev=start href=/ title=Home Page / which seems like it ought to be rel instead, there are typos in more cases than just rev=stylesheet. So the premise seems valid. -- Philip Taylor [EMAIL PROTECTED]
Re: [whatwg] Absent rev?
Philip Taylor wrote: rev=stylesheet makes up 57% of those uses of rev, How do you get that figure? even if you just compare rev=made(1157 instances) and rev=stylesheet(107 instances) you get 9.25% of the examples use rev incorrectly I will compare the rest of the results (if you like) but I cant imagine the figure will get any where near 57%? Thanks -- Martin McEvoy http://weborganics.co.uk/
Re: [whatwg] Absent rev?
... I think that ought to say ... (excluding rev=made, which is uninteresting since it's redundant with rel=author) In that case, rev is used on 0.2% of pages, which justifies the claim almost never used. And rev=stylesheet makes up 57% of those uses of rev, ... As far as I recall, a percentage of 0.2 usually is considered a big number, considering the total amount of pages. Or are we going to remove *every* feature that's used in less than 0.2% of the pages? BR, Julian
Re: [whatwg] Absent rev?
On Wed, Nov 19, 2008 at 2:54 PM, Martin McEvoy [EMAIL PROTECTED] wrote: Philip Taylor wrote: rev=stylesheet makes up 57% of those uses of rev, How do you get that figure? even if you just compare rev=made(1157 instances) and rev=stylesheet(107 instances) you get 9.25% of the examples use rev incorrectly That figure was from the case of ... (excluding rev=made, which is uninteresting since it's redundant with rel=author) since that appears to be what Hixie meant (but forgot to say) when claiming that most uses of rev were typos of rel. (Case-insensitively, I counted 1259 rev=made, 122 rev=stylesheet, and 1474 rev=... in total, which means 215 in total excluding rev=made, and 122/215=57%.) -- Philip Taylor [EMAIL PROTECTED]
Re: [whatwg] Absent rev?
Tab Atkins Jr. wrote: On Wed, Nov 19, 2008 at 9:26 AM, Philip Taylor [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] wrote: On Wed, Nov 19, 2008 at 2:54 PM, Martin McEvoy [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] wrote: Philip Taylor wrote: rev=stylesheet makes up 57% of those uses of rev, How do you get that figure? even if you just compare rev=made(1157 instances) and rev=stylesheet(107 instances) you get 9.25% of the examples use rev incorrectly That figure was from the case of ... (excluding rev=made, which is uninteresting since it's redundant with rel=author) since that appears to be what Hixie meant (but forgot to say) when claiming that most uses of rev were typos of rel. (Case-insensitively, I counted 1259 rev=made, 122 rev=stylesheet, and 1474 rev=... in total, which means 215 in total excluding rev=made, and 122/215=57%.) -- Philip Taylor [EMAIL PROTECTED] mailto:[EMAIL PROTECTED] In addition, a large proportion (looks like a majority, but I haven't explicitly calculated) of the remaining @rev showing up is rev=home, rev=back, rev=toc etc. which is clearly incorrect. Those people are assuming the @rev is meant to be a go back link, rather than just expressing a reverse-semantic version of @rel. (I highly doubt that these are links *from* home pages to inner pages, which would be necessary for the semantics to work correctly.) There are also a couple (3, it seems) of rev=shortcut icon, which is a similar typo to the rev=stylesheet one, and several rev=owns and similar which suffers from the same redundancy as rev=made (just replace it with rel=owner). So, by this survey, it looks like there's less than 50 correct and not-obviously-redundant uses of rev out of 127k, which puts it under 0.04%. ~TJ Here is my take on the subject. There are 1517 instances of @rev of those: made occurs 83% of the time (1259 instances) stylesheet occurs 8.2% of the time (124 instances) The rest occur 8.9% of the time (135 instances) the misuse of stylesheet is trivial and only a matter of informing authors of their error, the fact that a high amount of authors are using rev-made is Inspiring to say the least, because every made link type is a claim of ownership, not authorship two totally different semantics. I will study the results of @rel soon but from first glance It seems there is (statistically) more abuse and misunderstanding about @rel than there will ever be than @rev Thanks -- Martin McEvoy http://weborganics.co.uk/
Re: [whatwg] Absent rev?
On Wed, Nov 19, 2008 at 10:09 AM, Martin McEvoy [EMAIL PROTECTED]wrote: Here is my take on the subject. There are 1517 instances of @rev of those: made occurs 83% of the time (1259 instances) stylesheet occurs 8.2% of the time (124 instances) The rest occur 8.9% of the time (135 instances) the misuse of stylesheet is trivial and only a matter of informing authors of their error, the fact that a high amount of authors are using rev-made is Inspiring to say the least, because every made link type is a claim of ownership, not authorship two totally different semantics. I will study the results of @rel soon but from first glance It seems there is (statistically) more abuse and misunderstanding about @rel than there will ever be than @rev Well, at the bottom of this email[1] is the previous list with all the @rel removed. This way everyone can easily see the numbers without having to process the file themselves. As previously noted, there are 1266 instances of rev=made (y'all missed a few). (Note also that one of them is rev=made, publisher, which is impossible together short of AI.) There are 124 rev=stylesheet. There are 4 rev=shortcut icon, plus a few more obvious typos like rev=copyright, rev=nofollow, and rev=text/css (1 of each, so 3 total). The next group is somewhat subjectively wrong, so your own calculations may lead to some slightly different numbers. As well, it's possible I'm a few off in a category, so my numbers may not match up *exactly*. A few of these are mistakes where the author took the attribute as being short for revision. There are 3 of these. There are several being used merely to hold arbitrary data for scripting. There are 7 of these. There are many which are for whatever reason holding arbitrary non-scripting data, which can't be construed as anything but a confusion. I count 36 of these. Another large category are those which are *probably* intended to be @rel, but the author was confused about how @rel/@rev related and used it incorrectly. I count 29 of these. We also have several which fall into the same category as rev=made, where they are trivially redundant with an obvious @rel. (I define trivial here as can be turned into a @rel just by changing the conjugation of the word.) I count 26 of these. To be specific, the values I'm talking about here are owns, owner, author, designed, and powered. It's also possible that some of these actually fall into the previous category. Finally we come to the ones that I can't otherwise categorize. I'll just go ahead and list them directly: 3 a rev=vote-for ... 2 a rev=testfield ... 2 a rev=footnote ... 1 link rev=Subdocument ... 1 link rev=self ... 1 link rev=replyto ... 1 link rev=reply-to ... 1 link rev=child ... 1 a rev=vote-against ... 1 a rev=vote-abstain ... 1 a rev=review ... Some of these may fall into a previous category, but it's possible that they are being used correctly. Many of these also have a fairly trivial @rel that would express the same relationship as well. So, let's discuss the numbers. There about 1517 instances of @rev (my numbers add up to 1513, but as I said, there may be some mistakes in my numbers). 1292 of them are trivially reversible into an @rel. 160 of them *are* @rel values, but the author either typoed or misunderstood the @rel/@rev distinction. 46 of them are simply wrong, and don't have anything to do with the actual uses of @rel/@rev. 15 are possibly correct uses of @rev which aren't *trivially* reversible (that is, they actually require some thinking about the relationship to name the opposite relation for use in @rel). That makes 85% of the uses trivially reversible, 14% of the uses mistakes, and 1% of the uses valid and non-trivial to reverse into an @rel relation. @rev-using links are only 1.2% of the total body of links which use @rel or @rev. [1]: http://www.xanthir.com/rev_using_links20081119.txt ~TJ
Re: [whatwg] Absent rev?
On Wed, 19 Nov 2008, Martin McEvoy wrote: It basically says that the whole premise that HTML5 should drop *rev* (a) because authors use it wrong, (b) Many authors use rev-stylesheet wrong, is a MYTH and an inaccurate assessment of the *rev* attribute As others have noted, the data does in fact show that rev= is rarely used for anything other than rev=made, and is, with the exception of rev=made, usually used incorrectly when used at all. The idea of removing it is to make validators more able to report these mistakes, thus helping authors write better HTML. Despite your claims to the contrary, given the way that the rel attribute and the related keywords are defined, rel=author does in fact convey the semantics that rev=made did. Removing rev doesn't affect previous pages, as they continue to be valid HTML4 if they were valid HTML4 before, and UAs can continue to support those semantics for as long as they want to support those pages. Furthermore, since the definition of rel in HTML5 allows relationships in either direction to be defined, there is no need anymore for a separate rev= attribute. On Wed, 19 Nov 2008, Martin McEvoy wrote: There are 1517 instances of @rev of those: made occurs 83% of the time (1259 instances) stylesheet occurs 8.2% of the time (124 instances) The rest occur 8.9% of the time (135 instances) These numbers support removing rev= based on the design principles we are using for HTML5. the misuse of stylesheet is trivial and only a matter of informing authors of their error Well, who's going to be doing the informing? Nobody did it in the past ten years, why would they do it now? the fact that a high amount of authors are using rev-made is Inspiring to say the least, because every made link type is a claim of ownership, not authorship two totally different semantics. I believe it is unrealistic to expect authors to split semantics that finely. Authors who today use rev=made could equally well use rel=author without loss of generality IMHO. I will study the results of @rel soon but from first glance It seems there is (statistically) more abuse and misunderstanding about @rel than there will ever be than @rev That's possible, but we can remove rev= without reducing the semantics that can be expressed; we can't remove both without losing a feature. On Wed, 19 Nov 2008, Julian Reschke wrote: As far as I recall, a percentage of 0.2 usually is considered a big number, considering the total amount of pages. Most of hte 0.2% is rev=made, which is redundant with rel=author. Or are we going to remove *every* feature that's used in less than 0.2% of the pages? If there are redundant features that are only used 0.2% of the time, we should probably remove them, yes. Are there any? On Wed, 19 Nov 2008, Martin McEvoy wrote: That does not solve the problem of rev=made because its not the same as rel=author author can relate to multiple instances on a page saying WE made this, an Author may have no control over who claims authorship of a page. made is usually a single point perspective, Its a way of authors claiming their own links in a statement saying I made This. I don't understand your distinction. rev=made and rel=author are interchangeable, they would both appear on the same page (the page that was written), pointing to the page of the author. rev=made is subtle but ever so important link relationship for an author, HTML5 really shouldn't abandon rev because of it. While I appreciate your feedback, I'm afraid that in this instance the weight of the argument is more strongly in favour of dropping the attribute, thus it has been dropped. If you still disagree, please provide new information explaining why, for example explaining the use case for rev= that rel= doesn't address. -- Ian Hickson U+1047E)\._.,--,'``.fL http://ln.hixie.ch/ U+263A/, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
Re: [whatwg] Absent rev?
Ian Hickson wrote: On Wed, 19 Nov 2008, Martin McEvoy wrote: It basically says that the whole premise that HTML5 should drop *rev* (a) because authors use it wrong, (b) Many authors use rev-stylesheet wrong, is a MYTH and an inaccurate assessment of the *rev* attribute As others have noted, the data does in fact show that rev= is rarely used for anything other than rev=made, and is, with the exception of rev=made, usually used incorrectly when used at all. The idea of removing it is to make validators more able to report these mistakes, thus helping authors write better HTML. OK then... Despite your claims to the contrary, given the way that the rel attribute and the related keywords are defined, rel=author does in fact convey the semantics that rev=made did. No It doesn't Reverse and Inverse properties are key factors of any Semantics without both @rev and @rel there is hardly any semantics at all just a one way stream of information, which most of the time you have to guess what the Authors intentions were. rel=author on the whole only relates to published documents, rel=made relates to Documents, Music, Photos, Videos, Sunday Lunch! Literaly anything that can be *made* Removing rev doesn't affect previous pages, as they continue to be valid HTML4 if they were valid HTML4 before, and UAs can continue to support those semantics for as long as they want to support those pages. I cant see anyone abandoning HTML4 soon at least not in my lifetimebut you never know Furthermore, since the definition of rel in HTML5 allows relationships in either direction to be defined, there is no need anymore for a separate rev= attribute. So essentially @rel in html5 is breaking the semantics of @rel just because it cant deal with @rev? On Wed, 19 Nov 2008, Martin McEvoy wrote: There are 1517 instances of @rev of those: made occurs 83% of the time (1259 instances) stylesheet occurs 8.2% of the time (124 instances) The rest occur 8.9% of the time (135 instances) These numbers support removing rev= based on the design principles we are using for HTML5. the misuse of stylesheet is trivial and only a matter of informing authors of their error Well, who's going to be doing the informing? The publishers of HTML5 Nobody did it in the past ten years, why would they do it now? Nobody over the last 10 years informed Authors very about Validation and Accessibility, but they are at last getting to grips with it.. the fact that a high amount of authors are using rev-made is Inspiring to say the least, because every made link type is a claim of ownership, not authorship two totally different semantics. I believe it is unrealistic to expect authors to split semantics that finely. They do... Authors who today use rev=made could equally well use rel=author without loss of generality IMHO. OK then example: I am the author of numerous websites and I decide (like many people do) to place some links on my homepage a portfolio If you like. My Homepage is at : http://groovydeveloper.com/ Here is my link a rel=author href=http://somegroovysite.com/;Groovy Site/a Above Statement (In HTML4) says http://somegroovysite.com/ Authored http://groovydeveloper.com/ Which Is rubbish its the other way round The Same statement in HTML5 will say (because @rel is a reverse and inverse link type) http://somegroovysite.com/ Authored http://groovydeveloper.com/ and http://groovydeveloper.com/ Authored http://somegroovysite.com/ @rel seems to be redundant because describing the link with rel=author doesn't actually tell you who the author of a is page you have to guess, the statement is at most only half correct and again not expressing any real semantics [edits] If there are redundant features that are only used 0.2% of the time, we should probably remove them, yes. Are there any? A lot considering that the average website only uses 19 elements[1] How many are there in HTML5? [1] http://code.google.com/webstats/2005-12/pages.html On Wed, 19 Nov 2008, Martin McEvoy wrote: That does not solve the problem of rev=made because its not the same as rel=author author can relate to multiple instances on a page saying WE made this, an Author may have no control over who claims authorship of a page. made is usually a single point perspective, Its a way of authors claiming their own links in a statement saying I made This. I don't understand your distinction. rev=made and rel=author are interchangeable, No I guess you don't :-) While I appreciate your feedback, I'm afraid that in this instance the weight of the argument is more strongly in favour of dropping the attribute, thus it has been dropped. Unfairly From what I can tell Thanks for your help anyway -- Martin McEvoy http://weborganics.co.uk/
Re: [whatwg] Absent rev?
Martin McEvoy wrote: rel=author on the whole only relates to published documents, rel=made---oops! rev=made relates to Documents, Music, Photos, Videos, Sunday Lunch! Literaly anything that can be *made* But you knew that ;-) -- Martin McEvoy http://weborganics.co.uk/
Re: [whatwg] Absent rev?
Martin McEvoy writes: Ian Hickson wrote: given the way that the rel attribute and the related keywords are defined, rel=author does in fact convey the semantics that rev=made did. No It doesn't Yes it does; it's specified that they are equivalent: For historical reasons, user agents must also treat link, a, and area elements that have a rev attribute with the value made as having the author keyword specified as a link relationship. -- http://www.whatwg.org/specs/web-apps/current-work/multipage/structured-client-side-storage.html#link-type-author Reverse and Inverse properties are key factors of any Semantics OK. without both @rev and @rel there is hardly any semantics at all just a one way stream of information, That simply isn't true, since it's always possible to define rel=foo and rel=bar where bar conveys the same semantics as foo but in the opposite direction; no rev needed. which most of the time you have to guess what the Authors intentions were. Not at all, since part of the definition of the rel values says in which direction they are to be interpreted. For example, rel=author indicates that the referenced document provides further information about the author of the section that the element defining the hyperlink applies to. rel=author on the whole only relates to published documents, rel=made relates to Documents, Music, Photos, Videos, Sunday Lunch! Literaly anything that can be *made* So the problem is that if I wanted to be able to create a link from my Sunday lunch to its cook, annotating it as such, I wouldn't be able to do so because rel=author isn't appropriate terminology to use in meals? That's true, but given that my Sunday lunch isn't written in HTML anyway, I don't see how it could be trying to use rel=author (or indeed rev=made) in the first place! Ditto for all your other examples. By definition the thing which is making the rel=author link has to be written in HTML 5, and therefore has an author of some sort. Furthermore, since the definition of rel in HTML5 allows relationships in either direction to be defined, there is no need anymore for a separate rev= attribute. So essentially @rel in html5 is breaking the semantics of @rel just because it cant deal with @rev? No, the semantics of rel aren't changed from existing use; HTML 5 takes care not to break existing widespread use. the misuse of stylesheet is trivial and only a matter of informing authors of their error Well, who's going to be doing the informing? The publishers of HTML5 Why should they bother to do that when then can more easily define the problem no longer to exist? Authors who today use rev=made could equally well use rel=author without loss of generality IMHO. OK then example: I am the author of numerous websites and I decide (like many people do) to place some links on my homepage a portfolio If you like. My Homepage is at : http://groovydeveloper.com/ Here is my link a rel=author href=http://somegroovysite.com/; Groovy Site/a Above Statement (In HTML4) says http://somegroovysite.com/ Authored http://groovydeveloper.com/ Which Is rubbish its the other way round Indeed. The Same statement in HTML5 will say (because @rel is a reverse and inverse link type) I don't understand what you mean by the part in parentheses. Please could you expand on it, or provide a reference. http://somegroovysite.com/ Authored http://groovydeveloper.com/ and http://groovydeveloper.com/ Authored http://somegroovysite.com/ Of course not. See my quote from the rel=author part of the spec above; it clearly defines in which way that relationship applies. Among the set of relationships that rel allows there are relationships in each direction (both from and towards the current document), but a given relationship is always unambiguously defined to be in a particular direction. If there are redundant features that are only used 0.2% of the time, we should probably remove them, yes. Are there any? A lot considering that the average website only uses 19 elements[1] That simply doesn't follow. There are many ways in which hundreds of different elements could be distributed throughout a population such that each of them are used on more than 0.2% of pages yet the mean elements per page is 19. Smylers
Re: [whatwg] Absent rev?
On Wed, 19 Nov 2008, Martin McEvoy wrote: Despite your claims to the contrary, given the way that the rel attribute and the related keywords are defined, rel=author does in fact convey the semantics that rev=made did. No It doesn't Reverse and Inverse properties are key factors of any Semantics without both @rev and @rel there is hardly any semantics at all just a one way stream of information, which most of the time you have to guess what the Authors intentions were. rel=author on the whole only relates to published documents, rel=made relates to Documents, Music, Photos, Videos, Sunday Lunch! Literaly anything that can be *made* They are in fact _defined_ to be equivalent in HTML5: http://www.whatwg.org/specs/web-apps/current-work/#link-type-author I don't understand what benefit there would be to saying that HTML pages about photos couldn't use rel=author. That seems weird. either direction to be defined, there is no need anymore for a separate rev= attribute. So essentially @rel in html5 is breaking the semantics of @rel just because it cant deal with @rev? Could you provide an example of how rel's semantics are broken? the misuse of stylesheet is trivial and only a matter of informing authors of their error Well, who's going to be doing the informing? The publishers of HTML5 That would be me, and I assure you that I am not going to be doing any informing of the millions of authors who make this mistake. Nobody did it in the past ten years, why would they do it now? Nobody over the last 10 years informed Authors very about Validation and Accessibility, but they are at last getting to grips with it.. On the contrary, both the validation and accessibility efforts have spent massive amounts of resources on evangelisation. I believe it is unrealistic to expect authors to split semantics that finely. They do... The data suggests that the majority of authors do not distinguish subtle semantics like this. (I mean, more than 99% of people don't use rev= at all, for instance.) Authors who today use rev=made could equally well use rel=author without loss of generality IMHO. OK then example: I am the author of numerous websites and I decide (like many people do) to place some links on my homepage a portfolio If you like. My Homepage is at : http://groovydeveloper.com/ Here is my link a rel=author href=http://somegroovysite.com/;Groovy Site/a Above Statement (In HTML4) says http://somegroovysite.com/ Authored http://groovydeveloper.com/ Which Is rubbish its the other way round So say it the other way around, e.g.: pI wrote a href=http://somegroovysite.com/;Groovy Site/a./p You don't actually need a rel= at all. What problem is the rel= solving for you? If you really wanted to use rel=, you could define a new value, say sample-work, and use that: a rel=sample-work href=http://somegroovysite.com/;Groovy Site/a The Same statement in HTML5 will say (because @rel is a reverse and inverse link type) I don't know what you mean by reverse and inverse; where do the specifications define it that way and what does it mean? If there are redundant features that are only used 0.2% of the time, we should probably remove them, yes. Are there any? A lot considering that the average website only uses 19 elements[1] How many are there in HTML5? Many more; are any redundant? We've removed acronym because of it being redundant with abbr, I don't really know of any other redundant ones. On Wed, 19 Nov 2008, Martin McEvoy wrote: Martin McEvoy wrote: rel=author on the whole only relates to published documents, rel=made---oops! rev=made relates to Documents, Music, Photos, Videos, Sunday Lunch! Literaly anything that can be *made* But you knew that ;-) I believe this makes my point more strongly than anything else that has been said in this thread. Even someone who is asking for rev= to be kept (and thus can be assumed to be informed on the matter) makes the very mistake that our data shows is a common mistake. How can we expect your average HTML author, who couldn't care less about HTML, to get this right if even we get it wrong? -- Ian Hickson U+1047E)\._.,--,'``.fL http://ln.hixie.ch/ U+263A/, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
Re: [whatwg] Absent rev?
Ian Hickson wrote: [Lots of edits] I don't know what you mean by reverse and inverse; where do the specifications define it that way and what does it mean? @rev = This relates to That, or a rev=help link indicates that the current document is help for the resource indicated by the href. @rel = That relates to This, or a rel=help link indicates that the resource indicated by the href is help for the current document. Anyway I give up, this discussion is getting a little too testy, If you, And many others don't understand the point I am trying to make, what progress is there to be made, Its all just wasted time (something I don't have right now), Im sure HTML5 will be great for Browser Vendors, for the Humble author well we'll see. Thanks everyone for your...er...kind words see ya ;-) -- Martin McEvoy http://weborganics.co.uk/
Re: [whatwg] object element feedback
On Tue, 6 May 2008, Vlad Alexander (xhtml.com) wrote: Does the NPAPI define a way to submit form data? Yes, please see: https://bugzilla.mozilla.org/show_bug.cgi?id=188938 IE, FF and Opera support plug-ins to submit data in a form. And XStandard plug-in supports this. For example: object type=application/x-xstandard name=editor1 width=100% height=300 param name=Value value=Hello World / /object IE, FF and Opera will send editor1=Hello World to the server in an HTTP POST. I've added object to form submission in HTML5. There is a need for a cross-browser technique to install plug-ins. Until such a technique is specified, the codebase attribute should not be removed. It hasn't been removed, it just hasn't been defined yet. Not really sure how to define it. In IE it's a pointer to native windows code, in HTML4 it's a base URL for resolution of relative URLs... Not sure what to do with it. Most browsers seem to ignore it. -- Ian Hickson U+1047E)\._.,--,'``.fL http://ln.hixie.ch/ U+263A/, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'