Lists Home |
Date Index |
Weirdly or not, if HTML is still dominant, and HTML is an
SGML application, then "SGML on the web" is working or
at least, it's most successful application is.
XML isn't. Or if it is, it's biggest success is
in messaging and sending dataGlobs to the transform
engine and the local DOM data bucket.
Not exactly the results we want here. But not because
markup per se is failing. I think possibly because
people doing many things don't need it for very much.
1. SMIL: aka, powerpoint in markup. The problem
is that people have powerpoint and its toolset and
use that for web applications. (I like SMIL in
RealPlayer and it is a blast to work with but again,
2. VoiceXML. Neat but patent encumbered. It becomes
like Flash or PDF: it's owners have to push harder. XML
in and of itself has lots of good qualities but in no
way guarantees "openness".
3. XML Data islands: standard or not, they work.
4. XSLT: very powerful and much loved, but if one
is good at ASP scripting, one has a procedural mindset,
a rowset, and a way to transform. It isn't as good
as XSLT but 80% of the time, it's good enough.
In my experience, adopting generalized markup requires
a vision beyond interoperation and portability. One
has to be concerned about lifecycle as well and platform
independence. Still, see item 2, and realize that markup
doesn't guarantee platform independence or reusability except
from the point of view of XSLT, and data on the web doesn't
have a significant lifecycle yet. The value of the content
has to be high. That is why we used it for tech manuals.
Yes, the browser framework affects all of this significantly.
On the other hand, I'm not sure people will put up with
the extra work of using a full-up generalized markup browser
as we did in the days before HTML, or that it has much
meaning to them given that in such a beastie, one still
needs a stylesheet system that takes on all the semantics
of any conceivable rendering (eg, real time 3D) with
the attendant true extensibility issues where extensibility
means extend the framework itself.
So plugins and downtranslation are likely to be around for
some number of years to come.
Just for grins, consider what the classes for a framework
would be that could handle any conceivable markup application.
What would that look like? Who would be able to code for it?
From: Simon St.Laurent [mailto:firstname.lastname@example.org]
I've just posted "XML sur les navigateurs: XHTML, SVG, SMIL, XSLT et
plus" (in English except for the title) on my site:
Basically it takes a look at the ways "SGML for the Web" has (or,
largely, has not) affected the traditional browser-centered Web.
(I gave the presentation at AFNET's Net 2002 conference last week in
Paris - http://www.afnet.fr/afnet/net200x/programme.html#T9)