- From: Jonathan Robie <jonathan.robie@datadirect-technologies.com>
- Date: Wed, 18 Sep 2002 17:43:03 -0400
- To: www-xml-linking-comments@w3.org
- Cc: w3c-xml-query-wg@w3.org
The XML Query WG is pleased with the decision to refactor XPointer to allow very simple pointer schemes, without forcing every XML application that needs pointers to support the full complex hypertext system envisioned by the xpointer() scheme (which is not yet going to last call). Simple pointers are important for basic architectural reasons. On the web as a whole, we desperately need a *simple* way to point into XML documents. These pointers should be simple and efficient enough to be used in specifications like these: 1. XQuery needs to traverse pointers among documents. 2. XML Schema should extend its notion of referential integrity beyond the single document boundary. 3. RDF needs to identify nodes in a document. Ideally, the simplest form of pointer should not require DTD or schema processing, so we would actually prefer that the base XPointer framework not require DTD or schema processing. Pure "tumbler" syntax works great for this, as in this example using the element() scheme: element(/1/2) The need for a simple pointer along these lines has also been noted by Tim Bray and James Clark: ...maybe the #1 gaping architectural hole as regards XML & the Web. The problem is that at the moment, given some arbitrary XML, there is *no* good way to determine what's an ID without recourse to some external resource like a DTD or schema, and that, to use a technical term, sucks. Although we feel that the need to use external references to find IDs is an unfortunate aspect of the design, we note this issue has been discussed for years, and it is more important to have simple pointers than to fine tune this to get an even simpler form of pointer. We have to stop fiddling with this and get it out - the first Working Draft was April 6, 1997. We think that the simple schemes that are now going to Last Call are the ones that should be leveraged in most W3C architectures. The more specialized functionality of the xpointer() scheme is also important, but to a much smaller audience. That audience includes many scholars whose work should be supported -- commercial significance is not the only useful measure of value. However, the xpointer() scheme is too complex for server side architectures, and XQuery does not plan to support queries that traverse pointers defined with this scheme. Jonathan Robie On behalf of the Query WG
Received on Wednesday, 18 September 2002 17:43:40 UTC