[osis-core] Summary of discussion with Patrick regarding proposed conclusion for references.

[Patrick Durusau]

Troy,

What I am going to suggest to Todd later this morning is as follows:

For anything that you want to use in an osisID or osisRef, you have to 
declare a <work> element.

One of the parts of a <work> element is:

<identifier> which (will have) a workID attribute. That is a short form 
to refer to a longer identifier for the work. With the bibliographic 
information in the <work> element, you can fully identify any work that 
you want to use with the osisID or osisRef.

Now, that means that you can have:

<identifier workID="NIV"> for a work element that identifies the full 
publication data on the NIV

<identifier workID="Vug"> for a work element that identifies a 
particular edition of the Vulgate

All it is doing is declaring a value you can use later for 
osisID="Vug:Mark.1.1"

The part that I am going to suggest to Todd that we leave is the 
refSystem element in the  <work> element.

The reason for that is just like <creator> under <work> it is just 
another fact that is being recorded about the <work> that I am 
documenting. It does not have any meaning with regard to the citation 
that I am using in the text, although one would be advised to use the 
reference system of the work you are citing if you expect useful 
results. It in no way implies anything about any of the issues of 
transformation and the like that had Todd concerned.

This preserves your ability to record information about the work, such 
as the reference system that it uses, without making any claims about 
reference systems for the encoding.

Patrick

Troy A. Griffitts wrote:

> a few things.
>
> i feel in the dark about the <identifier> tag.
> there must be a way to express work:French.NIV refSys:KJV
>
>
> Todd Tillinghast wrote:
>
>> This is repost of the prior posting with an appropriate header.
>>
>> Following are a summary of a conversation Patrick and I had earlier
>> today regarding references.
>>
>> We came to the following suggestions/conclusions:
>> 1) Contrary to prior proposals by myself, there should be NO mechanism
>> or strategy to indicate that an identifier can be used without a
>> transformation for differing levels of the identifiers in the namespace.
>> External mapping strategies, techniques, software and/or specifications
>> will be used to determine if Bible.NIV:Matt.1.1 can "fall back" to
>> Bible:Matt.1.1.
>>
>> 2) <osisText> will have an attribute that self identifies itself as a
>> specific work (this attribute will likely be named osisWork).  This
>> value of this attribute must match the workID attribute of a <work>
>> element in its header.  This attribute will NOT define or declare the
>> default "namespace" to be used for identifiers in osisID or in osisRef.
>> A second attribute will declare the default "namespace" for use with
>> osisID identifiers and with osisRef.  The value of this attribute will
>> default to "Bible" and if a value is specified the value must match the
>> workID attribute of a <work> element in the <header>.
>>
>> 3) The <refSystem> element in <work> should be eliminated.  This removes
>> the confusion and ambiguity that exists regarding differences between
>> the <identifier> value which is a "namespace" in its own right and
>> <refSystem> with is also a "namespace".  As a result all references and
>> self-identifiers should be expressed in the context of a single
>> "namespace" as defined by the value of <identifier>.
>>
>> 4) There will be no mechanism to make a reference to one work using
>> another "namespace" and an identifier from that namespace.  The
>> identifier in a reference is always in the context of the "namespace"
>> and to the work that defines the "namespace". 
>> For example there will be no mechanism to make a reference to the text
>> in Bible.NIV based on the identifier Bible.French.LSV:Ps.42.13.  In
>> short there is no way to express Bible.French.LSV[Bible.NIV]:Ps.42.13
>> other than Bible.NIV:Ps.43.2.
>>
>>
>>
>> Examples:
>> Ex:
>> <osisText defaultNamespace="bible" osisWork="tev"> <!-- the
>> defaultNamespace could have been defaulted with the same result -->
>>     <header>
>>         <work workID="tev">
>>             <identifier type="OSIS">Bible.TEV</identifier>
>>         </work>
>>         <work workID="bible">
>>             <identifier type="OSIS">Bible</identifier>
>>         </work>
>>         <work workID="kjv">
>>             <identifier type="OSIS">Bible.KJV</identifier>
>>         </work>
>>         <work workID="fr">
>>             <identifier
>> type="OSIS">Bible.French.LSV</identifier>
>>         </work>
>>     </header>
>>
>>     <!-- ref to Ps.43.3 as defined in Bible.TEV and to the specific
>> work Bible.TEV -->
>>     <reference osisRef="tev:Ps.43.3"/>
>>     <!-- ref to Ps.43.3 as defined in Bible and to an abstract work
>> Bible -->   
>>     <reference osisRef="Ps.43.3"/>
>>
>>     <!-- ref to Ps.43.3 as defined in Bible and to an abstract work
>> Bible -->   
>>     <reference osisRef="bible:Ps.43.3"/>
>>
>>
>>     <!-- ref to Ps.42.13 as defined in Bible.French.LSV and to the
>> specific work Bible.French.LSV -->
>>     <reference osisRef="fr:Ps.42.13"/>
>>
>>
>>     <!-- ref to Ps.43.3 as defined in Bible.KJV and to the specific
>> work Bible.KJV -->
>>     <reference osisRef="kjv:Ps.43.3"/>
>>     ....
>>     <!-- same applies to osisIDs -->
>> </osisText>
>>
>>
>> Patrick, let me know if I have put any words in your mouth or left
>> anything out?
>>
>> Responses?
>>
>> Todd
>>
>>
>>
>
>

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