W3C home > Mailing lists > Public > public-evangelist@w3.org > October 2003

RE: WASP asks the W3 XHTML 1.0 or HTML 4.01

From: Brian Kelly <B.Kelly@ukoln.ac.uk>
Date: Thu, 30 Oct 2003 18:50:14 -0000
To: 'Jim Ley' <jim@jibbering.com>, public-evangelist@w3.org, molly@molly.com
Message-ID: <005a01c39f16$a7a5e570$d513268a@ukoln.ac.uk>

Jim and I had a similar discussion a few weeks ago.  This helped me
clarify my thinking, but I don't necessarily agreed with all of Jim's
views.

Anyway, in the context of the WASP document:

Why use XHTML 1.0?  

One possible reasons:

It's the latest recommended version of HTML, according to W3C. If this
is not the case (it's experimental, browsers aren't ready for it, etc.)
the W3C page should be updated to reflect this otherwise W3C are giving
the wrong signals.

The WaSP page says "XHTML is easier to maintain".

I would agree with Jim that this is misleading - the current generation
of tools aren't XHTML-ready; XHTML validators aren't widely deployed,
give ambiguous results, etc.

The WaSP page says 'The margin for errors in HTML is much broader than
in XHTML, where the rules are very clear.'

I'd draw the opposite conclusion.  The HTML spec allows mixed case for
elements, quoted or unquoted attributes in many cases, etc.  HTML is
therefore more tolerant of not only errors but also different approaches
to producing compliant HTML; unlike XHTML which does not provide such
flexibility.

The WaSP page says "HTML is easier to teach and to learn
The syntax rules defined by XML are far more consistent than those found
in HTML and therefore easier to explain than the SGML rules on which
HTML is based."

Partly true, but most people won't be learning from scratch and will
have to unlearn current HTML.

The WaSP page says 'XHTML is XSL ready'

I agree this is a valid reason - but there's still the MIME type issue
to consider.

Incidentally the MIME type issue is a reason why XHTML is not easier to
maintain.  Put an unescaped ampersand or an MS Windows character in a
HTML page and it's still viewable on many platforms; do the same in an
XHTML document with an XHTML MIME type and you get nothing.  XHTML is
less fault-tolerant but more rigorous - a simple "XHTML is easier to
maintain" slogan is quite clearly incorrect.

> 
> "Switching from HTML 4.01 to XHTML 1.0 brings almost no 
> direct benefits for the visitors of your Web site"
> 
> Almost no? - could you expand on the benefits it does bring, 
> as this was what I was hoping to get from the question, and 
> alluding to, but not listing the benefits has left me just 
> wanting more.

Wouldn't "Switching from HTML 4.01 to XHTML 1.0 currently brings almost
no 
direct benefits for the visitors of your Web site but the ability to
make use of XSLT provides the potential for a range of benefits" be a
reasonable comment to make?

> 
> I really do appreciate the work of the QA team and the WASP, 
> and I'm sorry that my responses to their articles are 
> generally negative, but I do believe their decision to defend 
> a bad technology decision (XHTML as text/html) should not be 
> supported, and it should have to actually talk the truth and 
> not gloss over the flaws.
> 
The position I'm coming to is that XHTML + correct MIME type should be
deployed today in circumstances in which compliant XHTML 1.0 can be
guaranteed - typically with software which can impose compliance doing
the creation.  XHTML 1.0 is not recommended for use with HTML authoring
tools unless they can force strict compliance or unless there is a
publication process which addresses these issues (and using vi to write
directly to Web file store can't do that!).

Brian

> Hi,
> 
> The first of my two almost certainly predictable responses to...
> 
> http://www.webstandards.org/learn/askw3c/oct2003.html
> 
> (This is for the errors/lack of clarirty in the document, the 
> other is to attempt to rebut the arguments)
> 
> [in XHTML 1.0]
> "empty elements are terminated using a space and a trailing slash"
> 
> no they're not in XHTML, they're closed by a trailing slash - 
> the whitespace is not relevant.
> 
> Next is not strictly an error, but in the context of 
> authoring XHTML 1.0 today, where Appendix C compliance is 
> essential, the example has:
> 
> <p xml:lang="fr"> ... </p>
> 
> This is not allowed under Appendix C. which requires that 
> xml:lang and lang be duplicated.
> 
> Also, you state that only syntax not semantics have changed, 
> I do not agree with this statement, in HTML and XHTML
> 
> <script type="appliction/x-jims-script">
> <!--
> // -->
> </script>
> 
> Has different semantics - in HTML the <!-- is part of a 
> script, in XHTML it is a comment.  The semantics may not 
> matter to most people, but the application/x-jims-script 
> actually considers <!-- to mean write the current date. and 
> // --> to write the current time.  Utterly contrived example 
> of course, but an example of how the semantics of elements 
> have changed IMO (what's a comment in XHTML is a script in HTML)
> 
> 
> "In HTML, [...] termination of many elements [...] are 
> allowed and commonplace."
> 
> Could you provide an example of an element in HTML which is 
> allowed to not be terminated. (as opposed to a tag of course 
> which can, AIUI it is not possible to have a valid HTML 
> document with elements not closed.)
> 
> Enough of the errors, although they need fixing before the 
> rest of the document can be taken seriously, now the actual 
> content of the document:
> 
> "Switching from HTML 4.01 to XHTML 1.0 brings almost no 
> direct benefits for the visitors of your Web site"
> 
> Almost no? - could you expand on the benefits it does bring, 
> as this was what I was hoping to get from the question, and 
> alluding to, but not listing the benefits has left me just 
> wanting more.
> 
> "XHTML is easier to maintain"
> 
> This glosses over the fact that there are no QA tools to 
> ensure XHTML Appendix C compliance, without these, I can't 
> see how the claim can be justified - especially as one of the 
> statements isn't even detectable by the W3's own XHTML 
> validator, or any other validator/QA tool I know of.
> 
> "The margin for errors in HTML is much broader than in XHTML, 
> where the rules are very clear"
> 
> What are unclear about the rules of HTML4 ?  The W3's 
> validator, and other SGML based validation appear to have no 
> problem in validating to the HTML rules - Is the document 
> intending to suggest that the HTML 4 specification
> is flawed?   I have large problems with the XHTML 
> specification for example
> Appendix C.1 suggests avoiding PI's and Appendix C.14 
> suggests using PI's - this contradiction in the specification 
> is not what I'd call clear.
> 
> I really do appreciate the work of the QA team and the WASP, 
> and I'm sorry that my responses to their articles are 
> generally negative, but I do believe their decision to defend 
> a bad technology decision (XHTML as text/html) should not be 
> supported, and it should have to actually talk the truth and 
> not gloss over the flaws.
> 
> Cheers,
> 
> Jim.
> 
> 
Received on Thursday, 30 October 2003 13:50:17 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Friday, 15 July 2011 00:13:22 GMT