- From: Simon St.Laurent <simonstl@simonstl.com>
- Date: Mon, 22 Mar 1999 13:17:15 -0500
- To: www-xml-linking-comments@w3.org, xlxp-dev@fsc.fujitsu.com
For the most part, I have fewer concerns with XPointers than I do with XLink, but the concerns I do have would probably be addressed by a fairly drastic step, the use of a 'Level' approach like that used in the CSS and DOM WGs. XPointers that refer to well-formed chunks of XML are easy to describe and easy to implement. This subset has enormous usage potential outside the fairly narrow set of tools used to support linking, and development using XPointer to reference objects generated from XML document processing is already underway. This subset is relatively uncontroversial, and would make a good 'Level 1'. The other set of XPointers are far more critical for use in text annotation and complex linking. As a developer whose hypertext interests derive from work with historical documents, I definitely want to be able to create links whose ends aren't well-formed, cross element boundaries at will, or may appear multiple times in the same document (keywords.) At the same time, however, I recognize that implementing XPointers whose targets are not simple elements is a difficult problem that can interfere with many of the tasks I described in the previous paragraph. As a result, I would suggest that these XPointers be described in a 'Level 2' document, making them available to developers who need this sophistication but without imposing on developers who only need 'Level 1.' While some may point out that partial implementations are something that happens all the time, 'official' recognition of this fact and the creation of a vocabulary to describe levels of XPointer implementation seems like a good way to keep developers happy while allowing users to find and take advantage of the tools they actually need. Simon St.Laurent XML: A Primer Sharing Bandwidth / Cookies http://www.simonstl.com
Received on Monday, 22 March 1999 13:15:40 UTC