[tei-council] constraint prose WINITA

[Syd Bauman]

Sourceforge is not being cooperative right now, so I can't file a
ticket, and I don't want to forget this.

The prose of <ref target="#TDTAGCONS">22.4.4.3 Additional
constraints</> needs some work. Mostly it says <gi>constraint</>
when it means <gi>constraintSpec</>. But also, e.g. "outside the
scope of conventional schema languages" asserts that rule-based
schema languages are not conventional, which their more ardent
adherents might wage war over.

Also, the 2 examples of non-Schematron <constraint>s are in the
tagdocs for <constraint> and <constraintSpec>, and for the former it
is the only example. Certainly the tagdoc for <constraint> should
show Schematron constraints. But shouldn't the main prose demonstrate
non-Schematron constraints?

Other nit-picks:

* I'm not happy with the value list of scheme= of <constraintSpec>;
  it forecloses the possibility of publicly available general-purpose
  constraint languages we haven't thought of. And why use "xsl" to
  mean "XSLT"? 

* There are no examples of <constraintSpec> (or <constraint> ... it's
  the same set of examples, of course) that demonstrate any content
  other than ( constraint ) or ( desc, constraint ). 

* What would an <altIdent> or <equiv> child of <constraintSpec> look
  like, and more importantly, what would it *mean*?

I'm planning to attack the obvious stuff myself and make tickets for
the more controversial stuff, probably Mon or Tue next week.