[osis-core] Eureka with respect to reference syntax!

[Todd Tillinghast]

I think the core problem we are having with references and the related
syntax are the following:

1) In most cases within most Bibles and most (if not all) references a
single user EVERY encounters a SINGLE reference system for Bibles will
exist.

2) It seems burdensome to force users to understand AND to specify a
reference system will almost certainly be constant for ALL users they
ever encounter.

3) If a user were to have to specify a reference system they would most
likely turn to a specific WORK to describe/define the reference system
they operate under.

4) There are only a hand full of Bible related reference systems (likely
two, possibly three) that 99% of users will ever come in contact with.
The remaining reference systems will be used by scholars.

5) A reference is ambiguous from an electronic perspective without
specifying the reference system.

6) Specifying a work to imply a reference system ties a reference that
would otherwise be independent of a work to a specific work
unnecessarily.

7) Since only the reference system is acting in the ROLE of a namespace,
it seems that we have possibly confused things by putting "work" on the
LEFT side of the ":".



BACKGROUND POINTS:
A) In the XML world the concept of a namespace is quite the norm as is
the notion of a default namespace.  Once people understand that there
can be other default reference systems (ie "confessions of Augustine",
still being unaware of the fact that alternate reference systems exist
for the Bible) the need for a prefix (that can be defaulted) to identify
the set of identifiers an identifier will make sense.   (Users will
understand that there can only be one default namespace and why
Bible.OSIS:Matt.1.1 is necessary when "confessions.augustine" is the
default reference system (aka namespace).

B) A reference without a work is meaning full on its own.  If we were to
be having a conversation and I were to mention "Matt.1.1", given the
context of our location we would both assume the reference system that
is predominant in the United States.  The reference I make to Matt.1.1
has no tie to any one specific work and I can use the reference to
identify a passage in a number of works.  If I had wanted to
specifically mention a specific version I would have said some thing
like "I found it interesting how Matt.1.1 reads in the Phillips".  

C) Since we are talking about electronically encoding references we must
provide a mechanism to make clear the reference system that would
otherwise be obvious in causal conversation.  Since electronic documents
and references do not carry a context such a context must be specified
to make the ambiguous clear.


PROPOSAL:
1) Define a reference system Bible.OSIS.

2) Make Bible.OSIS the default reference systems if no other reference
system is stated as the default.  (This makes things easier for the
people who don't know that an OSIS document could be used to encode
works other than the Bible, but is not necessary for the remainder of
the proposed points.)

3) Use the identifier prior to the ":" EXCLUSIVELY for the reference
system, and put the "work" if specified at the end of the reference.
This way if the reference system is defaulted there will be no ":" in
the normal XML fashon and there will be no confusion between the role of
"work" and "refSystem".

**********************************************************************
**********************************************************************
**********************************************************************
**********************************************************************
Regex form:
refSystem:canonical@grain-canonical@grain[work]
**********************************************************************
**********************************************************************
**********************************************************************
**********************************************************************

EXAMPLES:(default refererence system being Bible.OSIS)
Matt.1.1 equals Bible.OSIS:Matt.1.1
Matt.1.1[Bible.Phillips] equals Bible.OSIS:Matt.1.1[Bible.Phillips]

Examples: (default reference system being confessions.augustine) (my
apologies for my lack of knowledge of confessions.augustine)
Bible.OSIS:Matt.1.1
Bible.OSIS:Matt.1.1[Bible.Phillips]
x.1 equals confessions.augustine:x.1
x.1[confessions.augustine.XWork] equals
confessions.augustine:x.1[confissions.augustine.XWork]


If we don't want to default the reference system and give a bias to a
particular reference system as well as a preference to an individual
reference system targeted as a particular class of works, we can
"subclass" the core schema and specify a default reference system in the
derived schema.  This affords the same convenience to all users and does
not favor those with more influence over the schema creation.  (We could
also create a default in the core schema and provide derived schema for
the other cases.)


QUESTIONS:
Is this the proposed regex a better solution and do you think we should
adopt it?  If not why not?

Do we want to default the reference system?

Todd