xpath 2.0 comments

I guess people are getting their comments in now, I pretty much feel the
same as the rest of the community I think, that while xsl-t 2.0 is
reasonable, xpath 2.0 is awful, so far at least.
Stuff I don't like:
    1.  if, for expressions, leave that to scripting technologies, xsl-t
that use xpath. People seem to be expressing some trepidation on writing
xpaths in the proposed version, hey, I'm afraid of trying to debug
transforms, and it seems to me the ability to have conditional
expressions will increase the capacity to screw up a transform without a
very great enhancement of its abilities. 
Jeni Tennison remarked in an email some months back to me on the subject
that we could always hope that processors will give good error messages,
I personally have never been a great believer in the existence of good
error messages. 
    2. comments, I think this will just increase the illegibility of an
xpath, it's a pretty compact syntax and comments in the middle will do
the opposite of what people hope comments will, i.e improve legibility.
Envision comments in complicated regular expressions.
    3.  I've been a pain elsewhere talking against the tight integration
with xml schema, this seems to me especially wrong headed. But hell I
guess it's coming.
Stuff I like
  I'm writing this part because my mother always taught me to end on an
upbeat note, that said I'd be willing to do without the stuff I like if
I could get rid of the stuff I dislike.
    1. Sequences, ranges of sequences.
    2. some of the functions, although some of them I can't get at all,
like for example xf:Unique-ID() "assigned by the User" huh? That said,
excess functions could be problematic, will not gripe about any if I can
get rid of if, for expressions.

Received on Friday, 10 May 2002 16:21:07 UTC