Re: [i] XPointer and addressing in OHP-Nav
From: Jacco van Ossenbruggen (jrvosse_at_cs.vu.nl)
Date: Tue 21 Jul 1998 - 11:15:26 CDT
I think I largely agree with Petes previous mail. I'd like to add two other
- Should XLink/XPointer affect the work on OHP-Nav?
My anwser is "not at all!" as long as we make sure that we can, at least in
theory, develop wrappers and/or special locspecs to make OHP-Nav interoperable
with the XML stuff. This means that OHP-Nav must provide concepts that
abstract from XLink/Xpointer in a way that makes sure we can be interoperable
with XML related standards without becoming dependend on them. An actual
implementation of the wrappers would result in another neat demo for OHP-Nav
and could, as an added bonus, provide a good testbed for XLink/XPointer, which
brings me to the next point:
2) Should the work on OHP-Nav affect XLink/XPointer?
If we think there is something wrong with the current drafts, my answer is
"yes!". And as Lloyd already pointed out, _this_ is the time to do it:
- XPointer does not address linking into non-XML data. We can regard this as a
good scope limitation, and write XPointer wrappers for XML applications that
need them and simply ignore XPointer for all non-XML data. At the other hand,
Lloyd already pointed out that we may add stuff to XPointer that makes it more
applicable for non-XML docs. Sigi already warned us for trying to become
"HyTime-complete". But if we can come up with a "lite" set of extentions to
XPointer which makes the spec more useful in OHP-Nav, we should certainly try
to get this into the spec.
- There may, even in an XML-only environment, be other problems with XPointer
when we use in cobination with OHP-Nav. Is this the case? Sigi?
- An implementation of XLink/XPointer using OHP-Nav would provide another
convincing demo for OHP-Nav, a proof-of-concept implementation for
XMlink/XPointer, and a test environment for XLink and XPointer which could
reveal bugs in the spec in a very early stage. Can any of you make an estimate
how much work it is to write a wrapper which implements a small subset (e.g.
only adressing based on IDs and addressing on the basis of the occurance of a
string) of XPointer? And what about an XLink wrapper?
Jacco van Ossenbruggen Faculty of Mathematics and Computer Science
mailto:jrvosse_at_cs.vu.nl Vrije Universiteit, de Boelelaan 1081a
http://www.cs.vu.nl/~jrvosse/ 1081 HV Amsterdam (fax: +31 20 44 47653)
This archive was generated by hypermail 2.1.5
: Tue 13 Aug 2002 - 07:21:02 CDT