- From: Eve L. Maler <elm@east.sun.com>
- Date: Wed, 23 Feb 2000 14:59:43 -0500
- To: www-xml-linking-comments@w3.org
>Date: Wed, 23 Feb 2000 20:08:33 +0100 >From: Erik Wilde <netdret@dret.net> >X-Mailer: Mozilla 4.7 [en] (Win95; I) >X-Accept-Language: en-US,de-CH >To: "Eve L. Maler" <elm@East.Sun.COM> >CC: Steve DeRose <Steven_DeRose@brown.edu>, dorchard@ca.ibm.com, > bent@exemplary.net, David Lowe <dbl@eng.uts.edu.au> >Subject: Re: more thoughts about show=embed... > > > Argh. If the starting resource were a single node, I would want > > the entire location set described by the ending resource to be > > embedded in that one place. I'd be surprised at the results if > > the embedding behavior changed when the starting resource become > > a whole set of locations itself. > >that's a good remark. it is necessary to think about the expectations of >users, and also the irritating change in behavior when resources change >and therefore location sets change. > > > Also, while the pairwise correspondence makes logical sense, what > > about when the number of locations doesn't match up? > >that was my first thought when i read the remarks about "pairwise" >matching. and more seriously, because an xpointer addresses a location >*set*, it is not ordered, and therefore identifying *matching* locations >for pairwise ordering is impossible by definition... > >ok, so it sounds as if there is something to clarify. suppose i would >like to implement xlink right now, then i would have no idea how to >handle embed with location sets, and i would choose to handle it in any >way i like. a different implementor could have a different idea about >what should happen, and we would end up with non-interoperable >implementations, where the same set of resources produced a very >different result when viewed on both platforms. > >furthermore, the xlink processing model seems to be very brief. i can >imagine applications where i use xslt stylesheets attached to documents >that dynamically generate xlink content that then should be used for the >presentation of the resource (my dictionary example would be a great >application for this kind of xslt/xlink combination), and the same thing >could happen with embedded resources. how should something like that be >handled by a browser? > >cheers, > >dret. -- Eve Maler Sun Microsystems elm @ east.sun.com +1 781 442 3190
Received on Wednesday, 23 February 2000 14:58:38 UTC