[osis-core] osis_0108.3 Making milestone OSISID and OSISIDREF match. OPTION E!

[Todd Tillinghast]

> Todd,
> 
> With a big <snip> to cut to the chase (so to speak) ;-)
> 
> 
> >OPTION E:
> >An interesting side direction would be to do the following:
> ><verseStart OSISID="Bible.TEV..Gen.20.17-18"><altID
> >OSISID="Bible.TEV..Gen.17/><altID
> >OSISID="Bible.TEV..Gen.18/></verseStart>Because of what had happened
to
> >Sarah, Abraham's wife, the <divineName type="yhwh">LORD</divineName>
had
> >made it impossible for any woman in abimelech's palace to have
children.
> >So Abraham prayed for Abimelech, and God healed him. He also healed
his
> >wife and his slave women, so that they could have children.<verseEnd
> >OSISIDREF="Bible.TEV..Gen.20.17-18"/>
> >
> Just got up from a much needed nap but note that this solution takes
us
> out of the verseStart as milestone and to a verseStart as container
> (albeit of milestones). Certainly could be done that way, but what is
> the problem with:
> 
> <verseStart OSISID="Bible.TEV..Gen.20.17-18" CommonStart="Gen.17.17"
> CommonEnd="Gen.17.18"/>

There are cases where there are more than two references covered.
(Bible.TEV..Matt.1.2-6a)  In this case if some one is looking for
Matt.1.4 the would not find it.  And each attribute name must be
enumerated so we can't say there are zero or more common references.
That is the reason for using elements.  I don't think this will come up
that often so having it a little different looking in those cases will
not be that bad.

I can go with Option D but think Option E with ReferenceTypes rather
than OSISID my be the best option.
>Because of what had
> 
> where CommonStart and CommonEnd are validated against ReferenceType?
If
> you make them IDs (assuming on a different element as you suggest)
> doesn't that get you back into the situation where "more than one
> element can claim that as a legitimate ID?" type problem?
> 
I was thinking that in this case the "alternate" IDs would only be
claimed by a single milestone.  However it is possible (even likely with
the Matt.1.2-6a and Matt.1.6b-11 case for example) that that the same I
would be claimed twice. (Matt.1.6)  However if the "claimed" ids were
Matt.1.6a and Matt.1.6b then we would still be ok with OSISID.  This
goes to the question of why do we want IDs in the first place.  To
validate that milestones match?  To provide a solid/efficient pointing
place?  I am not that invested either way.  I'm afraid I want to change
to ReferenceType from OSISID for some reason in the future.  So I vote
for ReferenceType based on a gut feel.


> 
> 
> >
> >The beauty here is that the "expected" verse values would be present
as
> >OSISIDs (which will afford all of the benefits of ID to what would be
> >commonly expected) AND the main reference system still maintains its
> >integrity.
> >
> Yes, but remember all IDs are unique, document wide.
Yes, in a good thing in this case.  Remember that we expected the
milestone to claim the reference anyway.  
> 
> >
> >A variation on the same schema would be:
> ><verseStart OSISID="Bible.TEV..Gen.20.17-18"><altID
> >OSISID="Bible.OSIS..Gen.17/><altID
> >OSISID="Bible.OSIS..Gen.18/></verseStart>Because of what had happened
to
> >Sarah, Abraham's wife, the <divineName type="yhwh">LORD</divineName>
had
> >made it impossible for any woman in abimelech's palace to have
children.
> >So Abraham prayed for Abimelech, and God healed him. He also healed
his
> >wife and his slave women, so that they could have children.<verseEnd
> >OSISIDREF="Bible.TEV..Gen.20.17-18"/>
> >
> BTW, where is the extra .. coming from? Bible.OSIS..Gen.17 (as opposed
> to Bible.OSIS.Gen.17), two periods between OSIS and Gen as opposed to
one?

I tried to get response on that yesterday.  I was looking for a
delimiter that is valid in xs:ID that would indicate a separation
between the reference system and the reference itself.  (See the
previous email) This has a number of reasons.  Since we expect that the
reference system will be defaulted in some cases and not in others we
need to be able to compose a longer reference from the defaulted value
(See the email sent about an hour ago about a place to put the default
reference system).  In other cases we need to separate the reference
system from the reference.  We need to have both one and two name
reference system identifiers so we can't just say the the first
identifier is the reference system PLUS we would not be able to
differentiate between Gen the reference system and Gen when it is a part
of the default reference system.  (Bible.TEV, Bible.KJV,
TwentyThreeTales)

 I don't really like ".." myself but all of the other options I tried
were not valid for IDs.  (There are a lot of options for attribute
values but IDs are more restrictive.)  PLEASE SUGGEST A BETTER OPTION.  

> 
> Patrick
> 
> >
> --
> Patrick Durusau
> Director of Research and Development
> Society of Biblical Literature
> pdurusau@emory.edu
> 
>