[osis-core] schema 1.1_003 bug

[Steve DeRose]

At 04:35 PM -0400 08/20/02, Patrick Durusau wrote:
>Harry,
>
>Harry Plantinga wrote:
>
>>>Comments on my earlier post about merging refSystem/refSystems 
>>>with work/works?
>>>
>>
>>Well, I guess missed that question. I think that would make sense
>>as long as there were some optional element in the <work> element
>>to specify a reference system.  E.g.
>>
>><works>
>>  <work osisWork="conf">
>>    <refSystem>augustine_confessions.pusey</refSystem>
>>  </work>
>></works>
>>
>The question came up because the schema simply copied the structure 
>of work and works for refSystem and refSystems, the reasoning (such 
>as it was and this was during a conference) being that we needed the 
>same information for you to make a reference to a reference system 
>as we do work.
>
>But, thinking about it after Chris pointed to the obvious 
>duplication, what is the added value of having separate 
>bibliographic elements that both point to an internal ID?

The added value is that they are two conceptually different classes 
of information, and keeping them straight now makes it easier to keep 
them straight later.

At some point, one can imagine that entries in <works> would be 
locatable via some mapping mechanism that leads to a copy of the 
right document (say, a CEB NT or something). At that point, the 
entries in <refsystems> would lead to quite different objects: namely 
documents that define a reference system, provide means for mapping 
between reference systems, etc.

In the short term, we don't have those mechanisms in place; but if we 
collapse the two classes of identifiers then people who author 
documents now will not be able to set their documents up to work 
later.

Also in the short term, we could validate that the work portion of 
references points to a work, and the refsys part to a refsystem; that 
seems a good thing, if only to avoid accidentally requesting the 
"Hebrew" version with the "CEB reference system" when you meant the 
reverse.

>In other words, once you complete a <work> element, it has a 
>osisWork attribute. osisText has only one, but that does not mean 
>that other osisWork attributes, such as for osisRef cannot also be 
>in the document.
>
>Note that we did not reach the issue of how you declare a machine 
>processable (or otherwise processable) refsDecl, a la TEI. Not that 
>the refsDecl there is machine processable either, but that is 
>another issue.
>
>So, if we don't have a refsDecl, with the syntax to declare a 
>machine or otherwise processable syntax for references, doesn't an 
>osisWork attribute value fufill the same role? In other words, I am 
>using it to refer to augustine_confessions.pusey (as an osisWork 
>value).
>
>Ah, but if you want to explicitly declare a document's reference 
>system, perhaps you have some special linking or processing, hmmm, 
>wouldn't that be an attribute on osisText? Or do you need more 
>indirection than that?

Huh?

>
>Would you ever want to change the default reference system?
>
>I guess I am asking if a refSystem attribute on osisText would do 
>what you need or would more be required?

I don't see how such an attribute would help; but maybe I'm missing something.

>
>Certainly in 2.0, we need to do some machine processable refsDecl mechanism.
>
>Patrick
>
>>
>>At any rate, there should be a way of tying an internal identifier
>>such as "conf" to an elsewhere-defined reference system such
>>as augustine_confessions.pusey
>>
>>-Harry
>>
>
>--
>Patrick Durusau
>Director of Research and Development
>Society of Biblical Literature
>pdurusau@emory.edu


-- 

Steve DeRose -- http://www.stg.brown.edu/~sjd
Chair, Bible Technologies Group -- http://www.bibletechnologies.net
Email: sderose@speakeasy.net
Backup email: sjd@stg.brown.edu