[
Lists Home |
Date Index |
Thread Index
]
Pardon me for being confused and leading Paul down a garden path -- yes, it
is just another transformation, so there is no problem inside the style
sheet itself. However, there is a potential for confusion in the output
document, as the implementations I've seen accomplish the namespace
transformation by putting in an xmls declaration for the surrogate prefix
with the desired URI. Implementations would have to work a little harder to
canonicalize prefixes at the root element, given that namespace-alias can
introduce multiple prefixes for the same URI (the "natural" one bound to the
real URI in the original, and potentially multiple surrogate uses
transformed to the real URI).
-----Original Message-----
From: Joe English [mailto:jenglish@flightlab.com]
Sent: Thursday, August 29, 2002 4:57 PM
To: xml-dev@lists.xml.org
Subject: Re: [xml-dev] the infoset is two infosets (or even three?) [was:
Re: [xml-dev] linking, 80/20]
Paul Prescod wrote:
> David Carlisle wrote:
> >
> > namespace-alias does work at the level of uri, it aliases any elements
> > in one namespace to elements in another, whatever prefix is used on the
> > elements (you get no control over the prefix that is actually used in
> > the result) You use prefixes to specify the two uri, but that is the
> > same as the uses of prefixes in Xpath expressions, the prefix is just a
> > local reference to the namepsace uri, it is the uri that are used in all
> > later processing.
>
> I think that Leigh was pointing out that it pretty much requires you to
> have two prefixes for the same URI (especially the XSLT URI) which is
> not allowed by Joe English's proposal.
Actually, I don't think namespace-alias does introduce
any insanity (at least in the strict technical sense :-).
In the stylesheet, there are still two different namespace names
bound to two different prefixes. xsl:namespace-alias just
specifies an additional transform to be performed on output --
change the namespace name referenced by the stylesheet-prefix
to the one referenced by the result-prefix.
Identifying the namespace names by prefix instead of by URI
does violate the "no QNames-in-content" rule, sort of, but
none of the other Principles of Sanity.
> If prefixes were taken out of the
> mix entirely then the problem would go away. As you point out that is
> really a syntactic, not semantic, change.
--Joe English
jenglish@flightlab.com
-----------------------------------------------------------------
The xml-dev list is sponsored by XML.org <http://www.xml.org>, an
initiative of OASIS <http://www.oasis-open.org>
The list archives are at http://lists.xml.org/archives/xml-dev/
To subscribe or unsubscribe from this list use the subscription
manager: <http://lists.xml.org/ob/adm.pl>
|
