[tei-council] W3C Xpointer schema registry

[Syd Bauman]

[Cross-posted to Council and SO]


> Hold on there a minute. I hate to sound like a wet blanket, but I
> really don't think this is more urgent than completing the class
> work.

I disagree -- this is potentially *very* time sensitive. Maybe not,
but we have no way of knowing who else is queued up waiting to
register scheme names, and I don't think we should take the chance.
Turns out the W3C isn't open for business now, so I have to wait
until Monday.


> I am also a bit worried that we should be sending off proposals to
> the W3C, of any kind, which have not actually been seen by the
> COuncil as yet. Or were these "schemes" seen so long ago that I
> have forgotten them already?

Perhaps. They were first seen by the group as a whole in Ghent, I
think. But don't fret too much, even if you haven't committed them to
memory and written java implementations of them. If I understand
correctly, what we are registering with W3C is the name, not the
semantics. That is, we want to be first to the punch so that a
reference to the match() XPointer scheme means the TEI match()
XPointer scheme, not Joe's match() XPointer scheme. If we get there
first, Joe will be the one who needs to use a different name or use
the cumbersome namespace mechanism the XPointer Framework provides.

As a reminder to everyone, the names we're planning to register are

  xpath()
  left()
  right()
  range()
  string-range()
  match()

We had xpath2() in an earlier version of the list, but decided to
drop it because XPath 2.0 was only a draft at the time. Has it become
a real official W3C recommendation? If so, we should go ahead and add
xpath2() back to the list, no?