This is an archived snapshot of W3C's public bugzilla bug tracker, decommissioned in April 2019. Please see the home page for more details.
The XML Core WG just noticed that the latest (editor's) draft of SML 1.1 at http://dev.w3.org/cvsweb/~checkout~/2007/xml/sml/build/sml.html?content-type=text/html;%20charset=utf-8 has a normative reference to "XPointer xpointer() Scheme" http://www.w3.org/TR/xptr-xpointer/ It should be noted that this is a (non-last call) Working Draft from 2002 on which no work is planned. In fact, the xpointer() scheme failed CR, and was withdrawn as a W3C work item. You should not plan to reference the xpointer() scheme in your specification. paul Paul Grosso for the XML Core WG
I had been meaning to propose that instead of the xpointer scheme, the SML spec use the xpath1 scheme. I hadn't gotten around to it, but this seems an opportune place and time. As far as I can tell, there is no particular bit of functionality present in xpointer that is not present in xpath1, that SML depends upon. Indeed, at least some examples of the extra functionality (like being able to point at ranges that do not contain well-balanced XML) would probably be problematic for most SML implementations. So xpath1 is probably a better fit in any case. Of course, xpath2 should also be considered.
xpath1 and xpath2 schemes are not W3C recommendations. The element() scheme is a W3C recommendation but it is difficult to use and fragile since it uses numeric child sequences. E.g., element(/1/2) However, it is expressive enough to represent SML references since they always point to a unique element in a doc
I'm not quite sure what you mean. XPath 1.0 and XPath 2.0 are both W3C recommendations. It's true that no W3C Rec specifies their use as an XPointer scheme. But the problem which gave rise to this bug report is that the expression language of the xpointer() scheme is defined only by a Working Draft (and one on which no WG is now working); on that point, XPath 1.0 and XPath 2.0 are on a rather different footing. The element() scheme seems far too brittle to be useful for cross references in a dynamic system. Relying on it would make SML reference error prone, and at the same time ensure that most of the errors would not be readily detectable (because the reference will not fail to resolve, just resolve to the wrong element).
Although the xpath1 and xpath2 XPointer schemes are registered in the XPointer Registry ( http://www.w3.org/2005/04/xpointer-schemes/), they are not W3C recommendation (see see Paul Grosso's email http://lists.w3.org/Archives/Public/public-sml/2007Jul/0093.html) Sure XPath 1.0 and XPath 2.0 are W3C recommendation, but these should not be confused with the xpath1 and xpath2 XPointer schemes. The XPointer registry does not have a spec for the xpath2 scheme, and the spec for the xpath1 scheme (http://www.tei-c.org/P5/Guidelines/SA.html#SATSXP) is not a W3C recommendation
*** This bug has been marked as a duplicate of bug 4636 ***