Lists Home |
Date Index |
email@example.com (Eric van der Vlist) writes:
>> [David Megginson]
>> XML the way that it haunted RDF. If we start putting stuff this
>> esoteric into specs, we scare away people who want to use XML to do
>> things, however interesting the esoteric stuff may seem to us
I think the problem with RDF reification on that score is that it was
something that it looked like everyone needed to know in order to use
RDF. (Namespaces produce similar jitters on the XML side, I think, but
as long as we close our eyes and hold our nose and tell everyone around
us to do the same...)
I don't think Eric is proposing to make these issues part of XML 101.
> [Eric van der Vlist]
>Sorry if I have made all this more confused than it should be, but I
>have the feeling that with XPath, XML schema languages and now this
>XML serialisation of the infoset we are looping over variants of the
>same issue and reinvent a different wheel at each iteration and that
>it could be worth to spend some time to think about all this.
Yes. This loop seems to be spinning pretty rapidly at the moment, and I
don't see signs of it slowing down. While it's probably better for
developers who aren't designing generic XML tools and APIs to stay out
of this conversation, lest their minds join the loop, it's definitely
worth a closer examination by those already looping.
Speaking of which, general issues of applying tree-based tools to graphs
and the loops that can produce was a dominant issue at Extreme. XPath
for RDF/Topic Maps/XML Schema type systems/etc. came up constantly.
There are some genuinely interesting and clearly difficult issues
lurking here. It was impressive to see a number of deeply-thought
alternative solutions, but the problem space seems very large.
Ring around the content, a pocket full of brackets
Errors, errors, all fall down!
http://simonstl.com -- http://monasticxml.org