Re: Data Model(s) for XML 1.0 / XML Devcon / DOM / XSL / Query

[Ben Trafford <ben@legendary.org>]

At 05:29 PM 2/23/2001 +0000, Sean McGrath wrote:

> >Sean McGrath wrote:
> >
> > > isn't it time to accept that not specifying formal
> > > post-parse data model(s) for XML 1.0 was a big
> > > mistake?
>
>At 12:18 PM 2/23/01 -0500, W. E. Perry wrote:
>
> >In a word, no. Those post-parse plus post-additional-processing data
> >models are in effect being specified now by, among others, the very
> >groups whose work you cite here.
>
>Ah yes, but the after-the-fact, bifurcated struggle to hone in
>on post-parse data models has resulted in nasty impedance
>mismatches between some key, horizontal XML technologies
>that should be living in harmony but are not.

         <soapbox>
         A lot of these discussions (especially in regards to the evolution 
of XPath-based languages like XSLT and XQuery, the lack of cohesion between 
XML specs, interdependencies, etc.) sounds to me like: Why didn't you build 
the perfect thing the first time?

         It's fairly evident that XML technology needs to continue to 
mature. These sorts of discussions help immensely. However, I must confess 
that I'd like to see a lot more work like RDDL happen. That's a good 
example of where somebody complained about something, the list chewed on it 
for a while, then somebody got off their behind and made something happen 
to address the issue.

         What I'm hearing a lot on the list these days is much vapour, 
little substance.

         If people don't like something about XML, propose a solution, 
implement it, and make it available. RDDL has done that, to great response. 
Heck, we're even using it where I work. Just because W3C produces 
recommendations that get big company backing, etc., doesn't mean that it's 
the only place where significant XML work can be done. Note the success of 
SAX, which is not W3C-based.
         </soapbox>

         I agree with both Sean and Walter on this issue. In Walter's case, 
I agree that having a single, simple XML specification creates a basic 
framework from which we can develop interoperability. I agree with Sean 
that now is the time for a post-parse data model. One of the things that I 
am currently struggling with in my own work (to be made public any day now, 
I promise, just like I have been for the last year!) is getting linking 
information to survive document transformation. If the XSL and XLink WGs 
had done more work, earlier, to ratify something, this wouldn't be such a 
hairy mess. But they didn't, and in my opinion, a post-parse data model 
would have nicely handled that.

         Anybody got any suggestions for such a data model? I haven't been 
paying close attention to what's been going on out there...

--->Ben