W3C home > Mailing lists > Public > www-mobile@w3.org > August 2000

Re: Request for review: CC/PP specifications

From: Sean Palmer <wapdesign@wapdesign.org.uk>
Date: Thu, 31 Aug 2000 17:36:05 +0100
Message-ID: <000a01c01369$af668800$c3dd93c3@z5n9x1>
To: "Ian Stokes-Rees" <ian@decisionsoft.com>
Cc: "Johan Hjelm" <johan.hjelm@era-t.ericsson.se>, <www-mobile@w3.org>
Sir(s),
> We have done extensive work on the "tidy" package such that we are able
> to produce XHTML from any HTML document.
Any "near valid" HTML document. Tidy, excellent program 'though it is, still
can't handle anything too radical.

> And what disqualifies "Any
> system that does so will be based on proprietary means and fuzzy logic
> only" from still being a system that does what you have suggested is
> impossible?
It is impossible to do it perfectly every time because the majority of
people do not code HTML perfectly. XHTML by it's very nature should be valid
XML. HTML might work 99% of the time, but SGML is a funny thing. The spec.
for it is rather long...

> I am a bit confused by your use of the word "proprietary".
> Any HTML document has an implicit XHTML version that will render the
> identical output.
Yes, any well coded HTML document. But, for reasons discussed above, it wont
work all of the time.

> > There is no way that any other means than
> > XSLT will be anywhere near as effective or as well controlled. I would
be
> > happy to review an example of this system-in-motion, 'though.
> This is an incorrect statement.  You could easily use the XML modules in
> perl, or even better make use of XML Script ( http://www.xmlscript.org).
Perl isn't as useful as XSLT in my opinion. I could sketch out some
arguements for this, but i'll leave that for a later date.
XML Script, hmmm...all non-W3C recommendations. As this is a W3C
recommendation we are reviewing, I thought it might be better to debate with
W3C languages so that any results may be published in the next CC/PP draft.

> By saying "There is no way that any other means than XSLT will be
> anywhere near as effective or as well controlled" you are making a very
> dubious assumptions -- that XSLT is the perfect transformation language
> for XML to provide deterministic transformations.
True; XSLT falls down in many areas. However, at no point have I said it is
perfect, it is merely the best we have to go on at the moment.

> I think you will find
> that XML Script can perform every operation possible in XSLT, and do so
> in a _much_ more effective and well controlled manner.
I'm not up on XML Script, but i'll have a look at it and see what I find.
Can it be parsed server side?

Anyway, to sum up: My recommendation that the word HTML be changed to XHTML
was merely an indication that the Internet is moving towards XML languages,
and as such any future W3C recommendations should support XHTML rather than
HTML. The fact that CC/PP involves naught but XML and discusses WML
frequently (another XML language), lends support to this idea. HTML is
useful, but outdated, XHTML is a firm replacement. The bottom line is that
given a few months, it will be MUCH easier to convert XHTML to WML than it
will HTML to WML, although yes, HTML to WML is possible.
Of course, you could write any new garbage language in SGML and convert that
to WML too, but i'm not suggesting that anybody tries (or that HTML is
garbage)!

Kindest Regards,
Sean B. Palmer
President and Founder
WAP Tech Info - http://www.waptechinfo.com/
Received on Thursday, 31 August 2000 12:38:32 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:16:02 UTC