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

Re: ISSUE-4: Versioning, namespace URIs and MIME types ISSUE-60

From: Boris Zbarsky <bzbarsky@MIT.EDU>
Date: Wed, 18 Feb 2009 12:52:52 -0500
Message-ID: <499C4AF4.1010500@mit.edu>
To: Robert J Burns <rob@robburns.com>
CC: HTML WG <public-html@w3.org>
Robert J Burns wrote:

> What I am saying is 
> that authors should produce documents that are conforming and that they 
> should not re-implement an entire UA in javascript just to display an 
> image (this is good and sometimes overlooked advice for scripting in 
> general; let the browser be a browser).

Then question remains: how _should_ a site handle the situation where 
it's got an image element, a URI to show in it, and alternate text?

In particular, how should it handle this situation if it wants to have a 
basic library for doing various DOM manipulations so the various people 
doing the authoring don't have to reinvent the wheel all the time and if 
it has pages in both XHTML1 and XHTML2?

> The serialization issues already exist due to different capabilities of 
> XML and text/html parsers.

While true, it's possible to create an XHTML1-conformant image that 
serializes as text/html, while it's impossible to do so for an 
XHTML2-conformant image.

> This seems like another red herring 
> introduced into a discussion which is supposed to be about conflicts 
> (specifically name collisions that are not addressed by context), 
> between XHTML1, XHTML2 and XHTML5.

I don't see how it's a red herring to point out specific issues that are 
conflicting.

> And I would never suggest authors assume UA error handling will "do the 
> right thing", especially for an author trying to use a non-conforming 
> document

Please indicate to me how the document could possibly be made conforming 
here if all you have are the HTMLImageElement, the URI to show, and the 
alt text for the image.

> Once an author steps outside conforming authoring practices the need to test 
> against targeted UAs. Again, I don't know what these issues you're 
> raising have to do with the original topic of conversation.

See above.  The naming conflict means that it's impossible to author 
conformant content in some situations.

>> My concern is that an author who wishes to produce a conforming 
>> document via DOM APIs and include images in it might not be able to do 
>> so as the situation stands.
> 
> Well, we haven't found any conflicts for the conforming document 
> situation.

Again.  You have an HTMLImageElement, a URI, and some text that is 
alternate text for the image.  How do you put those together in a 
conforming way?

-Boris
Received on Wednesday, 18 February 2009 17:53:41 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 9 May 2012 00:16:31 GMT