W3C

Results of Questionnaire ISSUE-118 Specification breaks semantics of existing link relations "index" and "first" - Straw Poll for Objections

The results of this questionnaire are available to anybody.

This questionnaire was open from 2010-12-08 to 2010-12-16.

3 answers have been received.

Jump to results for question:

  1. Objections to the Change Proposal to Create a well founded consolidation of the link types
  2. Objections to the Change Proposal to Simplify the incumbent rel="" model
  3. Objections to the Change Proposal to Drop support for certain rel="" values

1. Objections to the Change Proposal to Create a well founded consolidation of the link types

We have a Change Proposal to change some link types to their pre-HTML5 meaning, but also change some other link relations to be synonyms and consolidate the set overall. If you have strong objections to adopting this Change Proposal, please state your objections below.

Keep in mind, you must actually state an objection, not merely cite someone else. If you feel that your objection has already been adequately addressed by someone else, then it is not necessary to repeat it.

Details

Responder Objections to the Change Proposal to Create a well founded consolidation of the link types
Toby Inkster No objections.
Leif Halvard Silli
Julian Reschke

2. Objections to the Change Proposal to Simplify the incumbent rel="" model

We have a Change Proposal to simply some link relations relative to HTML4. If you have strong objections to adopting this Change Proposal, please state your objections below.

Keep in mind, you must actually state an objection, not merely cite someone else. If you feel that your objection has already been adequately addressed by someone else, then it is not necessary to repeat it.

Details

Responder Objections to the Change Proposal to Simplify the incumbent rel="" model
Toby Inkster Strong objection.

This changes the definitions of rel values (especially rel=index) from HTML4, introducing differences between how it's defined in HTML5 and how it's defined virtually everywhere else (HTML4, XHTML 1.x, RDFa and the IANA link registry).

The proposal cites a single implementation as justification for the change. While I grant that this implementation is a major one, given that it's just one implementation, a small patch could bring it into line with the rest of the world. Wordpress has a history of security problems, so many of its users upgrade on a very frequent basis. Thus, if new releases of Wordpress were brought into line with the semantics of other implementations, data published using old releases would quickly diminish to insignificance.

Wordpress also has a history of using outputting fairly good, semantically correct HTML. Thus if clear guidelines on which rel values are synonymous, which differ, and how they all should be used were included in the HTML5 specification, such as the guidelines summarised in the first change proposal, it seems likely that Wordpress' developers would be amenable to accepting such a patch.
Leif Halvard Silli I object strongly:

This proposal will lead to lots of confusion. HTML5 will then contradict both specs and implementaitons, which will only be confusing and which therefore will not have the intended effect. There is no strong belief in the semantics suggested in this proposal. Instead, specs and implementors have largely worked in the same direction towards something else that is more credible.

I also do not believe that the browser that are lacking a site navigation bar (Firefox, Safari, Chrome, IE) are likely to implement such a thing because some features are removed/concatenated into fewer features: the medicine for getting more and broader suppore simply isn't the right one - it will not work. What use is there in having fewer, less useful, items on the site navigation bar? The medicine suggested in CP 1 is much more likely to increase implementations!
Julian Reschke This proposal breaks the semantics of "index" badly; compare the definition in HTML4 (which *is* in use) to the one in the spec.

3. Objections to the Change Proposal to Drop support for certain rel="" values

We have a Change Proposal to drop support for certain rel entirely, based on lack of interest from users and implementors. If you have strong objections to adopting this Change Proposal, please state your objections below.

Keep in mind, you must actually state an objection, not merely cite someone else. If you feel that your objection has already been adequately addressed by someone else, then it is not necessary to repeat it.

Details

Responder Objections to the Change Proposal to Drop support for certain rel="" values
Toby Inkster Weak objection.

These semantics seem useful enough to keep in the specification.
Leif Halvard Silli I object strongly.

(1) I object to evalutating these link relation keywords based on the HTML5 editor's interpretation of them: When this CP (CP3) suggest to remove "index, up, first, last", then it also incororporates removal of the synonyms (according CP2) of these rel values: "begin, start, top, contents, toc, end". As the CP says, we will only have prev and next left.

(2) is questionable what effect such a removal from HTML5 will have. Today there are two major vendors which support these features: Opera and Mozilla. (Mozilla legally counts as the vendor of SeaMonkey, see http://weblogs.mozillazine.org/seamonkey/archives/2010/03/see_no_monkey_d ) Some WG members from Opera have said that they are interested in removing it. I have not picked up that SeaMonkey developers are interested doing the same thing. And removal from HTML5 does not constitute an obligation to remove support - there are other flavors of of HTML and XHTML (HTML32, HTML4, XHML1.x XHTML+RDFa etc) which keep these features.

(3) HTML5's outline algorithm is an expression of "page wide" semantics, which in turn can be used to create a TOC for the current page — see The Semantic Notepad page of the IE9 Test Drive site — http://ie.microsoft.com/testdrive/HTML5/SemanticNotepad/Default.html
However, (a) authors are also interested expression the (hierarchical) relationship between pages — pages can contain a TOC, can contain a index. In that regard SeaMonkey supports *more* features than those which are found inside HTML5: it has suppor for chapters, sections and subsections. (I recommend visiting this page with SeaMonkey and look inside its Site Navigation Bar: https://developer.mozilla.org/devnews/) With the maturity that SeaMonkey has w.r.t. to these features, it I don't seem them removing them just because HTML5 says so. Rather, it seems to me that HTML5's outline features will lead to an *increased* interested in hierarchic/navigational link relations.

(4) Several browser extensions exists (and have existed) for these features. Currently the browser exension market is rapidly increasing. Thus we could soon see more extension implementations rather than less. Extension implementations should be counted in - what the "major browsers" do natively is not the only thing that matters.
Julian Reschke These four relations are already in the IANA registry, so the overall effect could be ok depending on the outcome of ISSUE-127 (the location for the registry). However, the *current* impact would be that these link relations would become invalid, and I have to object to that given their wide use.

(Reminder: use of unregistered link relations currently makes a document non-conforming, a constraint *not* being checked by validator.nu).

More details on responses

  • Toby Inkster: last responded on 9, December 2010 at 11:47 (UTC)
  • Leif Halvard Silli: last responded on 11, December 2010 at 17:02 (UTC)
  • Julian Reschke: last responded on 13, December 2010 at 21:38 (UTC)

Everybody has responded to this questionnaire.


Compact view of the results / list of email addresses of the responders

WBS home / Questionnaires / WG questionnaires / Answer this questionnaire