Re: Removal (Time for XMail?)

That's probably the preferred method -- the principle purpose of namespaces
in general is as a mechanism for scoping spaces, and an a:id|b:id
nomenclature would work well here. This is also a place where the decision
to drop id() from XSLT makes great sense, because I can see this causing all
kinds of problems in a namespace context.

-- Kurt Cagle
----- Original Message -----
From: "Michael Brennan" <Michael_Brennan@Allegis.com>
To: <xml-dist-app@w3.org>; "'James Snell'" <jmsnell@intesolv.com>
Sent: Friday, September 29, 2000 11:43 PM
Subject: RE: Removal (Time for XMail?)


> > From: James Snell [mailto:jmsnell@intesolv.com]
> > [...]
> > Most definitely... one possible solution that I've starting
> > thinking about
> > was some kind of scoped-id mechanism.  I'm not sure how this
> > would work
> > exactly so it is more of just a random thought, but it would
> > be nice to be
> > able to declare ID's within a particular scope in the
> > document itself.  In
> > others words:
> >
> >   <doc>
> >     <some_data id='a1'/>
> >     <some_data id='a2'/>
> >     <some_data id='a3' local_scoped_id='1'>
> >        <item id='a1'/>
> >        <item id='a2'/>
> >        <item id='a3'/>
> >     </some_data>
> >   </doc>
>
> This is an interesting problem that had not occurred to me before. Could
one
> do the scoping using namespaces? When using some sort of generalized
> enveloping scheme, one would undoubtedly use namespaces, as SOAP does (or
so
> I would think).
>
> So to cite a similar example to yours:
>
>   <a:doc xmlns:a="some-URI" xmlns:b="another-URI">
>      <a:some_data a:id='a1'/>
>      <a:some_data a:id='a2'/>
>      <a:some_data a:id='a3'>
>         <b:item b:id='a1'/>
>         <b:item b:id='a2'/>
>         <b:item b:id='a3'/>
>      </a:some_data>
>    </a:doc>
>
> Then one could reference id "a:a1" versus "b:a1". One could even use a
> namespace just for the id attributes. Would this be a reasonable approach?
>
>

Received on Friday, 29 September 2000 20:33:24 UTC