W3C home > Mailing lists > Public > public-html@w3.org > February 2009

Re: HTML is a declarative mark-up language

From: Boris Zbarsky <bzbarsky@MIT.EDU>
Date: Sat, 31 Jan 2009 22:00:11 -0500
Message-ID: <4985103B.2020600@mit.edu>
To: Robert J Burns <rob@robburns.com>
CC: HTML WG <public-html@w3.org>

Robert J Burns wrote:
> The situation I'm mainly thinking about is the one I've raised before. 
> An author has a phrase that the author wants to serve as an anchor 
> destination. There is no other reason the author would markup the phrase 
> except because it needs to serve as an anchor destination.

Ah, ok.  I'm still at a bit of a loss to think of such situations where 
there is not something more appropriate to mark up the phrase with, but 
I can see it in principle, I guess.

>>  <a id="mytable"></a>
>>  <table><tr><td>
>>    This is an important table that people should link to
>>  </td></tr></table>
...
> I'm not sure what that has to do with the present conversation. 
> Obviously this is a pattern we need to support in implementation 
> conformance due to legacy content.

Yes, of course.

> However, the present conversation is 
> about whether we should allow authors to continue to use <a> in HTML5 
> document conformance, so I'm not sure what this example has to do with 
> that topic.

No, the present conversation at this point seems to be about whether <a> 
should be given special semantic status as a link target (or anchor 
destination; the two terms seem synonymous to me).  That's what you're 
suggesting the meaning of an <a> with an id but no href should be, right?

> To address the issue you raise in terms of document 
> conformance, I would think we should say this is a deprecated usage of 
> the <a> element in HTML5 that instead the author should wrap the table 
> in the <a> element use the id attribute on the table itself (if the 
> author deemed the anchor as unnecessary).

That might be acceptable, yes.

> I'm not following your logic. What are the above-described reasons that 
> would make <a id='anid' >an anchor destination</a> a bad idea for 
> document conformance.

The observed over-reliance on use of <a> where semantically an id on an 
existing node is more correct.  While this is an authoring issue, I 
believe it arises as a result of the special status <a> has historically 
had as an anchor destination, and I don't think we want to perpetuate 
this particular author perception.  That is, we should make it very 
clear to authors that you do NOT need an <a> to have an anchor destination.

>> Sadly also vice versa in some cases...  The meaning of some elements 
>> is very difficult to describe, much less understand, without reference 
>> to the processing model...
> 
> Not in terms of normative language however. It helps authors to 
> understand document conformance norms by informatively notifying them of 
> presentation examples, however, can you come up with an example where 
> the authors need to be told specific implementation norms to understand 
> the document conformance norms.

Since the conformance norms include when it is and is not appropriate to 
use the element (that is, the semantics of the element are generally 
considered to be part of the conformance norms from what I can see), it 
seems to me that a full understanding of the element's effect may in 
fact be needed to understand whether it can be used.

-Boris
Received on Sunday, 1 February 2009 03:01:02 UTC

This archive was generated by hypermail 2.3.1 : Monday, 29 September 2014 09:39:01 UTC