Re: handling namespaces in advance Re: [xsl] namespace required in transform

Subject: Re: handling namespaces in advance Re: [xsl] namespace required in transform
From: "James Fuller" <james.fuller@xxxxxxxxxx>
Date: Sun, 21 Jul 2002 13:22:37 +0100
----- Original Message -----
From: "Michael Kay" <michael.h.kay@xxxxxxxxxxxx

> I don't think it's ever likely that XSLT or XPath will be extended to
> handle XML that isn't namespace-well-formed. Even the raw XML 1.0 spec
> warns people not to use colons that way.

Hello Mike,

I think your statement might be de-emphasising my main point, yes it is true
that people misuse colons, and thats a side matter, and I agree that
xslt/xpath or whatver will never be extended to handle anything other then
well formed xml. The rest of the question is directed at the xslt list as a
whole....

I don't see why the actual syntatic construction of xml could not itself be
abstracted( instead of angle brackets, why not slashes...etc ) and defined,
a sorta schema for base/physical format; maybe this is a route of
introducing binary xml....ok yes there are issues all over the place, just a
thought.

back to my original meaderings;

The issue I have is one of performance in a streaming environment, imagine a
multi node SOAP server process, There may be at any one time, a few servers
talking to each other in the form of generating a HTTP request with xml body
or a HTTP response. Another analogous situation would be very long SAX
processing.

one could introduce possibly <test:test></test:test xmlns:test="urn:test">
at the end of processing...., could force the processing model to wait until
all data is gathered, just another thought.

In any event I dont see any harm in wanting to handling unknown namespaces,
why for instance does an 'error' have to throw if a namespace is unknown,
why not assign a temporary unique id and handle it, effectively giving us
the hook to handle it.

cheers, jim fuller




 XSL-List info and archive:  http://www.mulberrytech.com/xsl/xsl-list


Current Thread