- From: Joseph Reagle <reagle@w3.org>
- Date: Fri, 24 Jan 2003 12:48:54 -0500
- To: Paul Grosso <pgrosso@arbortext.com>, David Carlisle <davidc@nag.co.uk>
- Cc: spec-prod@w3.org
On Friday 24 January 2003 12:30, Paul Grosso wrote:
> Schema-validity should, of course, be doable.
I expect at some point we'll have XHTML+XInclude. I believe that's a feature
of XHTML2.0, but also expect there might be an intermediary
identifier/namespace/DTD from XHTML1.* .
> This is what XInclude's parse="text" attribute is for. Quoting [1]:
> ...
> I'm not sure what xpointer support would have to do with this.
> When you use parse="text", the "resource is treated as plain text"
> so xpointer is irrelevant.
The tricky bit is actually between these two requirements. I know a common
source of errata in my specs is typos between the in-line and external
schema and in the in-line examples. Consequently, it'd be better to
maintain the schema and examples externally, and keep them all valid.
(Which means not in little fragments.) So, when I want to include part of
an example or schema I want to use xptr to select the relevant part and
include it. If I understand, I can't do that if the parse type is text; and
if it's not text then it is returned as XML. What I'd need it a:
parse="xml" return="text"?
So in my script I evaluate the XPath, Canonicalize the result,and return the
encoded XML (e.g., '<', '>','&') in a <pre> element:
nodes = expression.evaluate(context)
...
for node in nodes:
chunk = Canonicalize(node,unsuppressedPrefixes=[])
chunk = '<pre class="%s">%s</pre>' % (hclass, encode(chunk))
Received on Friday, 24 January 2003 12:49:02 UTC