discussion proposal for bug 5395

During today's meeting it was suggested that I should produce some
sample wording illustrating what I meant when I suggested that the
definition of 'schemaComplete' should ideally be more declarative
and less imperative than in some other possible formulations.

I attach a sample showing several possible changes to the spec:
introduction of the term 'SML-IF model validation', some additions
to the informal description to mention the definition of that
concept, and a very rough approximation to the new section in 5.2
that Ginny suggested would be the right place to put the definition
of the meaning of schemaComplete.

In the interests of allowing concrete discussion, I have included
three possible descriptions of the meaning of schemaComplete:
one using the MUST NOT wording, one using a more declarative
wording, and one using the declarative wording and adding the
additional content we discussed today as 'level 2' (the idea that
schemaComplete = true entails an assertion that the schema 'has
what is needed' for full validation).

Additionally, I have added two forms of a non-normative note intended
to make clearer some of the consequences of the definition.  The
first form of the note is intended to clarify the consequences of
today's 'level 1'; the second form goes with 'level 2'.

--CMSMcQ