...then please consider RUNNING FOR YOUR LIFE FIRST. If you haven't touched it, or have only used it casually, you can simply have no idea what horrors, trolls, dragons, balrogs, orcs, etc are to be found in the darkest recesses of that spec. I spent four hours in a meeting this morning with seven people all of which can be considered to have good to excellent understandings of XML Schema working out issues with just two little fragments of the beast. Very scary things were continuously uncovered throughout the meeting.
It's a good thing we'll have full support for Relax NG soon...
The best thing that the W3C can do is recognize that James Clark writes excellent software and excellent specs. XSLT 1.0 has a few wrinkles, but it works and it works well because (1) James has a lot of experience in the field and (2) wrote an implementation concurrent with the specification. Neither can be said for the WXS WG.
Now, if someone else could follow that model (instead of the W3C's standard bureaucratic and bloated method of spec writing), then perhaps big parts of the XML foundation (like Schemata) might be close to usable.
Re:Ayup.
darobin on 2003-01-03T17:14:06
I very much agree overall except for the last paragraph. As a member in a WG (SVG), I can't say that I find the process to be bureacratic or bloated. On the contrary, it's rather open and creative. I think that depends a lot on a WG's internal politics, as well as on external pressure to get something. No one believed in XPath, XSLT, SVG, etc. but people had been expecting WXS for ages and wanted everything in it.
In an evil way, I can't wait to see what WSX 2.0 will look like
<g>Re:Ayup.
ziggy on 2003-01-03T19:01:28
The W3C is a big place, and some WGs are bound to be better than others. With my recent experiences with WXS 1.0, and the drafts for XQuery 1.0/XSLT 2.0/XPath 2.0, I have a big case of The Fear when it comes to the W3C process. The process that gave us WXS 1.0, XQuery et. al. is substantially divorced from the James Clark model: brilliant people implementing and specifying and using concurrently. That has two major drawbacks: vocabulary design that's exceedingly vague and hard to use/read, and a spec that's difficult to implement correctly.I do see your point though: the SVG WG has respected the holy triad of develop/use/specify concurrently.
Re:Hmm...
darobin on 2003-01-06T09:32:55
I'd say that's most likely the path that walks closest to sanity
:)
