- From: Barb Munsil <barbara.munsil@internetMCI.com>
- Date: Tue, 25 Aug 1998 23:33:23 -0600
- To: xsl-editors@w3.org
- Cc: wes_munsil@csgsystems.com
Thank you. I see that now. I apologize for my error, and for mistyping the Cc address. James Clark wrote: > *[attribute(foo)] does NOT mean the same as attribute(foo). The latter > returns the foo attribute of the current node, whereas the former > returns the current node's element children that have a foo attribute. > > Barb Munsil wrote: > > > > Dear Sirs: > > > > The effect of the production for NodePatterns is to allow the final ElementPattern in an > > ElementPatterns, if it is the form *[attribute(QName)], to be replaced by the shorthand > > attribute(QName). This allowance is non-orthogonal in two respects: it is only made for the final > > ElementPattern; and it is not made for any other Qualifier or combination of Qualifiers. Moreover, > > it seems to be of little practical utility. The grammar and language would be simplified, with no > > loss of expressive power, if it were removed. This could be done, for example, by removing the > > production for NodePatterns, and replacing NodePatterns by ElementPatterns in the production for > > AncestryPattern. > > > > Thank you for your consideration. > > > > Dr. Wes Munsil
Received on Wednesday, 26 August 1998 01:34:02 UTC