Lists Home |
Date Index |
On Thu, 2002-05-09 at 12:35, Matthew Gertner wrote:
> Cool, I was about to respond to your other post, but this is a much
> clearer formulation of the issue. Your initial premise is exactly
> right: what gets developers excited about XML is the prospect of
> schema-enabling it. Certainly that's true for me (your humble PSVI
> poster boy), and you're also right that I don't like CDATA, NOTATIONs
> and the like.
> I think where you are missing the boat is the assertion that somehow
> there could be some alternative representation of the PSVI that
> wouldn't be XML but would satisfy all the gearheads out there. Frankly
> this is a crazy idea. It took decades for something like XML to
> appear, and it's a huge boon. We are developing software right now
> that uses XML along with schema (having created our own version of the
> PSVI two years ago), and we also use XML parsers, XML editors, XPath,
> XSLT and a whole slew of other XML technologies and tools. Why on
> earth would we reinvent the wheel when XML works for us!? Just to
> preserve the "purity" of the language for the benefit of some markup
I think I was arguing something much stronger, actually. I was arguing
that XML may in fact be _bad_ for PSVI-style information exchange. The
lack of a popular XML format for representing the PSVI is just one
Sure, XML has the tools, but the tools don't require an XML
representation of information beyond the parser. XML 1.0 is a pretty
ugly foundation for building information interchange systems, though it
certainly broke through a logjam.
I asking whether the "next phase of XML" really needs to be bound to
XML, or whether XML's lexical baggage is just holding the PSVI work
back. "Something like XML" is a good idea for information interchange.
I'm not sure that text-based markup is really what programmers are
I'll let most of the rest of your post go, since I don't think my own
motivations for suggesting this have much to do with whether it's useful
> Maybe you are being purposely provocative (I can relate, lord knows),
> but the idea that XML+schema is somehow no longer in the spirit of XML
> is absurd.
I don't think it's such a hard case to make. The types native to XML
1.0 are element types and a few bits for attributes. That's a very
different set of stuff from the types native to W3C XML Schema.
Ring around the content, a pocket full of brackets
Errors, errors, all fall down!