Re: ISSUE-54: doctype-legacy-compat

Philip TAYLOR (Ret'd) wrote:

> The 2.0 "Strictly conforming" example then goes on to include :
>> <html xmlns="" xml:lang="en"
>>       xmlns:xsi=""
>>       xsi:schemaLocation="
>>                 "
> so would I be safe in assuming that when you write
>> So, if there should be version specified, then it should definitively
>> live in attribute or element value, probably version attribute would be
>> the best choice then.
> you are advocating something along the lines of the <html> element
> + attributes cited immediately above ?  

No. xsi:schemaLocation holds location of schema and doesn't carry
version information. Actually specifying schema location in instance
documents is very bad practice.

> If so, I agree that this
> would allow an HTML document to unequivocally define to which
> version of the specification it is claiming adherence, in which
> case I would also agree that duplicating this information in
> the DOCTYPE is not essential and may well be contra-indicated.

If version should be carried, then it should look like

for HTML serialization:
<!DOCTYPE html>
<html version="5.0">

and for XML serialization.
<html version="5.0" xmlns="">

> I would also be interested to know (are there any XHTML 2.0 WG
> members here ?) what benefit(s) the XHTML WG perceive in this
> duplication.

I don't think that mandatory DOCTYPE and xsi:schemaLocation can survive
to final version of XHTML 2.0 (if there is ever such spec published).
Using those features is considered bad practice between XML experts. And
please note that the latest XHTML 2.0 WD is more then 3 years old. I
think that XHTML WG agreed on changing those ridiculous conformance
criteria but since that time no new draft was published.


  Jirka Kosek      e-mail:
       Professional XML consulting and training services
  DocBook customization, custom XSLT/XSL-FO document processing
 OASIS DocBook TC member, W3C Invited Expert, ISO JTC1/SC34 member

Received on Tuesday, 20 January 2009 21:41:13 UTC