- From: Dave Pawson <dave.pawson@virgin.net>
- Date: Tue, 2 May 2000 20:07:48 +0100
- To: <www-dom-xpath@w3.org>
Francis wrote: So am I right in saying that what you want to do is track your user's navigation of some kind of tree-structured XML display, both to navigate it as easily as possible at run-time, and also because you can build an xPath expression that can be kept as a bookmark for later re-use. And that without xPath you would have to do more work to track your user's tree-navigation, and you would also have to design, implement and debug a way of storing that string of DOM navigation steps in such a way that it could be re-played later? The use case per se is an end user with (or without) a pc. the most complex is a student at a pc whose preferred media is audio. A dom view gives a cursor in sync on the text with the audio being spoken from a .wav file of the xml source. sync is via SMIL + a navigation centre (linking audio and xml). The student says, whoa, go back. Key combo (say C-B) used. 'Back' like this has to be interpreted wrt to context. preceding:: *[1] strikes me as a fair guess. Any comments? Similarly other 'goto' type instructions. UP is a good one. XPATH offers ancestor::*[1] should take me to a node at which I can then start 'reading' from its first child having text(). Place bookmark 'here' is my reason for wanting 'generate-id()' from XSLT. Any student wants to annotate a book. I'll need a notepad and a link to N1045 (or something similar) such that 'tomorrow' when the student comes back to the book (hangover or not :-) he/she can find the 'last place I was reading' ... and what did I say I had to do today? xpointer comes in when xpath leaves off. Highlight a block of text (C- C-e) then annotate that. xpointer has the capability to provide the two end points, whereas xpath can only 'point' with a granularity of element. I'm not sure I would need to store a string of navigation steps, but if you put yourself in the shoes of the student, you can imagine what you might want to do with paper surely? HTH DaveP Sounds to me like a pretty compelling argument that not having xPath available would require you to re-invent most of the xPath wheel. > > Have you not seent the one that's been posted to xml-dev? > http://www.246.ne.jp/~kamiya/pub/omquery.zip > I hadn't loaded it up - there is also (or was - no response right now) XPathDOM from http://www.falconwing.com/~schen/, and, as Julian Reschke points out, the selectNodes() method in MSXML. Oracle also appears to have gone with selectNodes().
Received on Tuesday, 2 May 2000 15:08:16 UTC