[osis-core] Linguistic Annotation Module Design Document - schema extension.

[Todd Tillinghast]

Kirk and Steve,

If you were to define enumerated _type_ values in a _base_ linguistic
schema but put NO enumerated values in the _base_ and then extend the
_base_ for each language and enumerate the values specific to the given
language:
1) You would not have to release a new schema for each new language
added
2) You could validate values with the schema
3) It would be easier for users to encode by not tripping over what will
become very long lists of possible values
4) If there are different data models and extensions on a per language
basis those subtle extensions could be made.

Todd