- From: John Cowan <cowan@mercury.ccil.org>
- Date: Thu, 16 Aug 2012 14:58:17 -0400
- To: David Lee <David.Lee@marklogic.com>
- Cc: Uche Ogbuji <uche@ogbuji.net>, "public-microxml@w3.org" <public-microxml@w3.org>
David Lee scripsit: > Will the same xpaths return the same result set as say an uXML xpath > processor ? In particular will > > //*[xml:id = "abc123"] > > or the equivalent xslt match expression ... be the same in an XML, > XML+NS , uXML processor ? I think so for XPath 1.0, provided you don't try to use the local-name() function. I'm not familiar enough with XPath 2.0 to say for sure. > If that passes the Turing test then I think we can claim a reasonable > data model compatibility. But more important to me a *user > expectation* compatibility. +1 > I am fairly certain that by completely eliminating colons then the > proposal so far would work for all simple names. But will it work > with the above ? I think it might but not sure. How could it? If you ban colons in names (option #1), "xml:id" can't be a MicroXML name at all. That's the situation I want to avoid. I'm willing to give up arbitrary prefixes because of the pain they induce when trying to keep XML+namespace compatibility, but that doesn't apply to the "xml:" prefix. -- Principles. You can't say A is John Cowan <cowan@ccil.org> made of B or vice versa. All mass http://www.ccil.org/~cowan is interaction. --Richard Feynman
Received on Thursday, 16 August 2012 18:58:40 UTC