- From: mozer <xmlizer@gmail.com>
- Date: Sat, 11 Aug 2007 20:28:36 +0200
- To: "Vasil Rangelov" <boen.robot@gmail.com>
- Cc: public-xml-processing-model-comments@w3.org
On 8/11/07, Vasil Rangelov <boen.robot@gmail.com> wrote: > > >Hopefully Norm will add it to the agenda and we'll see what the rest of > >the WG thinks. > > Thanks. > > >I think the argument for the status quo is probably that only elements > >and attributes can/should hold data: that it would be bad practice to > >have a markup language in which you'd *need* to match text nodes, > >comments or PIs. For example, for matching text nodes to be useful, > >you'd need a design like: > > > > <links>index.html<sep />archive.html</links> > > > >which isn't something we'd want to encourage. > > Well, here's an argument against that. Hope it helps. > > Sometimes, the input XML may not be under the direct control of the pipeline > author and (s)he may have to process exactly the bad type of XML you're > showing. True, the pipeline author can always create an XSLT transformation > to process that into something more XProc friendly, but if expanding the > URIs is all that (s)he wants, then the process might get inefficient as > there has to be one transformation for original-to-XProc, > p:make-absolute-uris, and another transformation for XProc-to-original. may be in the case you're talking about, you just want <p:string-replace match="links" replace="string()"/> <p:make-absolute-uris match="links" /> So it's may be not worth adding complexity to make-absolute-uris Regards, Xmlizer
Received on Saturday, 11 August 2007 18:28:40 UTC