[osis-core] Re: OSIS_0104 (internal beta)

[Todd Tillinghast]

Notes on OSIS_0104

(Do you guys prefere the numbered individual email issues or them as a
bunch?)
(If you all [Troy, Steve, Patrick, and Chris] are getting this message from
osis-core then let me know and I will not send mail to you individually as
well as to osis-core and save you the headache of duplicate email.)

1) Should <year> be derived from xs:year or xs:year or xs:gYearMonth or
g:date so that it can actually be validated.  These are accepted xml data
types why not use them.  The reason to use xs:YearMonth is if different
copyright can be given in the same year.  I don't know that much about
copyright.  In any case it seems that xs:year is more appropiate then
xs:string.

2) <revisionDesc> can now all kinds of things as decendants via <p>.  This
sort of behavior is not intended and is just the result of <p>s behavior.
Should we NOT use <p> by reference in this case an just create a simple <p>
"inline" within <revisionDesc>.

3) The comments in titleGroup about <title> being mixed, etc.. are no longer
true.

4) No generic milestone?

5) Why not have the attributes "startNote" and "endNote" in <note> be of
type reference?  Also is would seem that there would need to be a "work"
attribute needed to make these work in any case.  (Same comment for
<figure>)

6) <back> is still not optional.

7) If <publisher> and <pubPlace> are going to be required we should also
require (or at least force a default) type where the types for <publisher>
would be something like "print", "electronic", etc.. and the types for
<pubPlace> would be something like "physicalLocation", "electronicURI",
etc..

8) Type for <author> should be required (or at least defaulted).  It seem
there was a long list of authorTypes in the previous version.  Was it
deleted by mistake?

9) <date> should be derived from xs:dateTime!  (Not from xs:date to allow
for time to be used by electronic publishers.)

Todd