W3C home > Mailing lists > Public > public-xhtml2@w3.org > March 2007

Re: XHTML Basic 1.1 - changes list, examples, test suite

From: olivier Thereaux <ot@w3.org>
Date: Thu, 29 Mar 2007 23:46:05 +0900
Message-Id: <172A8101-4807-48B4-BC31-3897BF1BFC9E@w3.org>
Cc: "HTML WG" <w3c-html-wg@w3.org>, "XHTML WG" <public-xhtml2@w3.org>
To: "Steven Pemberton" <steven.pemberton@cwi.nl>

Hi Steven, thank you very much for your answer!

On Mar 29, 2007, at 22:48 , Steven Pemberton wrote:
>> I have now added that document type in our catalog, may I request  
>> that:
>> * You inform us (www-validator) when there is any new document  
>> type created by the HTML WG
> It's not clear to us why you need to add it to the catalog. There  
> is a public URL for the sysid. Our experience is that Basic 1.1  
> works out of the box with the validator.

Two reasons why adding a new doctype to our catalogue is a good thing:

* A catalogue provides a (crude, but effective) caching solution.
   sysids are great, but I don't want validation to cause countless  
requests to the DTD files.

* take the example of validating an XHTML (Basic 1.1) document served  
as text/html
   - the media type is not enough to determine unambiguously whether  
to parse as XML or SGML
   - there may not be an xml declaration
   - the doctype is not known in the catalogue
Without any unambiguous way to know it's handling an instance of an  
XML language, a parser/validator is likely (supposed?) to parse the  
document, and the DTD it refers to, as SGML. That's a recipe for  
disaster, as you can imagine. Having the doctype for any new xhtml  
language that can be served as text/html helps us avoid that kind of  
... It also lets us tell the validator what the "human readable" name  
of XHTML Basic 1.1, so that people can get a nice "This is valid  
XHTML Basic 1.1", which is better than "this document is valid -// 
W3C//DTD XHTML Basic 1.1//EN"
... It also lets us tell the validator that for this document type,  
what mime type is allowed, what mime type is preferred.

So maybe perfect XHTML Basic 1.1, served properly as application/xhtml 
+xml will work "fine" with the validator, but the reason we need a  
validator is that there are plenty of documents that are not perfect.  
For those, giving the validator intimate knowledge of document types  
helps greatly.

>> * You inform us of any publication or republication of DTDs, so  
>> that we can update the catalog of the validator?
> That information is always in the W3C news, wouldn't that be enough?

As I wrote:
>> as it will not only help us stay in sync with your specs, and help  
>> adoption of the specs, it will also expose an important audience  
>> to your work on these specs.

In other words, it would make sure people working on validators (all  
of them) know about the new language, and can test, update their code  
and catalogues, etc. If there are people or groups whom you think  
should know about your work, assuming that they are closely following  
broadcasted news is not the most effective way, not to mention  
probably a bit hubristic. On the other hand, contacting people  
directly make them feel like they matter to you, and you make sure  
they get the message. IMHO, at least, good direct communication has a  
low cost, and high benefits :). YMMV.

Maybe someone in the working group would be interested in providing  
that kind of liaison?

>> I would also strongly suggest adding to the XHTML Basic 1.1 spec:
>> * a list of changes from XHTML Basic 1.0, which it replaces
> Of course; that has already been done. Please see the abstract.

Ah, thanks. If my experience is any indicator, however, I would  
suggest that the abstract is not a good location for the list of  
changes. A separate section, accessible for the TOC, was what I (and  
others I expect) was looking for.

My original message was also mentioning examples in the spec, is that  
on your roadmap?

Thank you.
Received on Thursday, 29 March 2007 14:46:12 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:30:27 UTC