From: Clark C. E. <cc...@cl...> - 2004-09-02 15:43:47
|
On Thu, Sep 02, 2004 at 08:08:55AM -0700, Sean O'Dell wrote: | So you think the way XML refers to and loads its schemas is | dead-wrong? That somehow they failed in that regard? Because XML | schemas are very widely used, and they're delivered primarily over | http using URLs to both identify and locate them. What I'm proposing | for YAML is no different than that. This is a false assertion. The XML namespaces specification [1] does not describe any such "schema loading" mechanism from XML namespaces, neither does the XML schema specification. In fact, the XML namespaces specification goes out-of-its-way to state that an XML namespace treats IRIs "as strings, and they are identical if and only if the strings are identical". For concrete example, the namespace for XHTML 1.0 (probably the most prolific XML namespace) is http://www.w3.org/1999/xhtml, which gives you an HTML page, not a schema discovery mechanism. The same is true with just about every other 'http' based XML namespace; they do not have automated ways to automagically refer-to and load the corresponding 'schema'. In fact, beacuse people were in disagreement as to which schema was at the end of the rainbow, people invented RDDL. However, http://www.w3.org/1999/xhtml is _not_ an RDDL page, is it? If you want to persist with this line of argument, at least provide me a serious system that has been widely adopted that does this sort of thing. It's been many years since XML came out... so I'm sure if it was a good idea, and as obviouly right as you suggest, you won't have any problem showing me how to do this with _existing_ tools in just about every parser, etc. To quote the namespace FAQ [2]: Q: Can I resolve the URI as a namespace name? A: Yes. You can also eat a tractor, but that doesn't mean it is a good idea. Clark [1] http://www.w3.org/TR/xml-names11/ [2] http://www.rpbourret.com/xml/NamespacesFAQ.htm#s12 |