Lists Home |
Date Index |
Derek Denny-Brown wrote:
>>From: Michael Kay [mailto:email@example.com]
>>It's too late to fix the problems now, but it's easy to identify them.
> Agreed. Understanding the problem means we can build APIs which
> alleviate the worst pain for the average user.
This sounds like arguments that were used for XSD or SOAP-RPC, or
EJB, a few years back, or some I still hear for WS specs from time
to time. In other areas, as technologists, we'd rip and replace in
heartbeart; not doing so where it's actually needed - I'm baffled.
Also, I'm reluctant to depend on APIs to protect me from bad
underlying tech - that's the kind of morass XML normally helps me
stay away from. Sometimes things need to be fixed at the root.
I'd be interested inseeing what we could throw together here. Last
time around the spec highlight was IRI support -surely we can do
> Writing a new spec
> doesn't those users. Starting work on a new, cleaner design isn't a bad
> idea, but we also need to address the concerns of people writing
> software today. It will take another 5 years to get that spec
> implemented as ubiquitously as XML is now.
After spending about that length of time whining about them, I am
glad the the suspension of technical judgement on XML Namespaces
seems to be coming to an end here.
Anyway ubiquity isn't for much if that spectrum of tools are as
inconsistent as you mention ;)