[tei-council] class updates: some post-Kyoto progress

[Syd Bauman]

> I plan to look into moving <addSpan> and <delSpan> into att.spanning
> later this weekend.

Done. However, using classes here means that the spanTo= attribute is
optional, whereas previously the to= attribute had been required.
While I realize that we have not yet formally approved the use of
Schematron constraints in P5, I went ahead and inserted Schematron
rules to ensure that spanTo= is present.

* I'm wondering if It would be better if I could modify the attribute
  after it had been inherited from att.spanning, i.e. have
       <attDef ident="spanTo" mode="change" usage="req"/>
  in the <elementSpec>s for <addSpan> and <delSpan>. Pipe dreaming, I
  suppose. 

* This means a pair of extra error messages from `make valid` because
  when jing does NRL validation it doesn't permit elements from other
  namespaces. (I have no idea why not; does anyone?)