[osis-core] RE: Work Issue

Todd Tillinghast osis-core@bibletechnologieswg.org
Thu, 30 Jan 2003 14:41:48 -0700 

(Patrick, the last 5 paragraphs relate to you post.)
Troy,

You are correct. I the statement should be all of the sub-elements
except the <refSystem> element.

Ex.
 
The reference in this document
<osisText osisWorkID="somethingelse">
	<header>
		<work osisWork="nasb">
			<title>NASB</title>
			<identifier type="ISBN">1234567890</identifier>
			<refSystem>Bible.KJV</refSystem>
		</work>
		<work osisWork="somethingelse">
		</work>
	</header>
	<div>
		<reference osisRefs="nasb:Ps.1.2"/>
	</div>
</osisText>

would be a reference to 

<osisText osisWorkID="nasb">
	<header>
		<work osisWork="nasb">
			<title>NASB</title>
			<subjext>Bible</subject>
			<identifier type="ISBN">1234567890</identifier>
			<identifier type="OSIS">Bible.NASB</identifier>
<refSystem>Bible.NASB</refSystem>
		</work>
	</header>
	<div>
		<reference osisRefs="nasb:Ps.1.2"/>
	</div>
</osisText>

but would not be a reference to 

<osisText osisWorkID="nasb">
	<header>
		<work osisWork="nasb">
			<title>NASB</title>
			<subjext>Bible</subject>
			<identifier type="OSIS">Bible.NASB</identifier>
<refSystem>Bible.NASB</refSystem>
		</work>
	</header>
	<div>
		<reference osisRefs="nasb:Ps.1.2"/>
	</div>
</osisText>

because the sub-element "<identifier
type="ISBN">1234567890</identifier>" is not present.

The fact that the <refSystem> that was used to form the reference is
different than the <refSystem> in the target document is, as Patrick
points out, a mapping problem to be left up to software.

However, related to Patrick's post, we have defined that an osisID can
have identifiers from more than one reference system.  When a single
reference system is used exclusively within osisIDs, then the current
system of using "osisWork" to identify the THIS <work> element with its
single reference system works fine.  

When more than one reference system is used, the current system forces
the definition of two work elements to represent THIS document (one for
each prefix used in osisIDs).  If we are going to use multiple reference
systems within the current document we must allow multiple <refSystem>
elements within the THIS <work> element.  But if there are multiple
<refSystem> elements then use of the "osisWork" attribute as a prefix
becomes ambiguous, and requires that an identifier tied to a specific
<refSystem> element be used as the prefix that identifies not only which
reference system but also the THIS <work> element.

The reason there are not any texts marked up with multiple reference
systems is not a function of the lack of its usefulness or necessity,
but a symptom of the fact that few texts have been marked up.  I believe
that Troy is the man behind most (if not all) Bibles that have been
marked up.  In Troy's case, (Troy correct me if I am wrong) the source
for his encodings is oriented completely around the KJV reference
system, and thus all of his output reflects that.  However, we know that
several Bibles have different variations on the KJV reference system,
and I would expect that a number of Bibles would be encoded with both
the KJV and the translation specific variation.  Further I see not
reason why there would not be encodings that purely support more than
one different reference system (KJV, Hebrew, French). 

If you believe that the prefix for identifiers in osisIDs NEED NOT be
related to the THIS <work> element then this discussion would take on a
different flavor related to what those other <work> elements mean.  This
may be where we are on different pages.

Todd


-----Original Message-----
From: osis-core-admin@bibletechnologieswg.org
[mailto:osis-core-admin@bibletechnologieswg.org] On Behalf Of Troy A.
Griffitts
Sent: Thursday, January 30, 2003 1:13 PM
To: osis-core@bibletechnologieswg.org
Subject: Re: [osis-core] RE: Work Issue

I think the misunderstanding here is seen in the statement by Todd:

"In Toronto we determined that when resolving a reference that ALL of 
the elements in the <work> element in the referring document must be 
present in the THIS <work> element of the document being referred to."

First, I'm a little confused by this statement.  What elements are you 
talking about?  This doesn't make any sense if you really mean ALL. 
Maybe you mean just all the <identifier> subelements?

<refSystem> is one subelement that definitely would not fit in this 
rule-- in my opinion, unless I'm missing something.

<refSystem> merely defines the default reference system that WE'RE using

in regard to this work.  In the THIS work element, it specifies my 
default ref system (in both my self-identifying osisID's and also any 
self-referencing osisRef's); in other work elements, it specifies the 
default ref system I'M using when targetting this work with an osisRef.

I'm still not sure I understand the problem.

	-Troy.



Patrick Durusau wrote:
> Todd,
> 
> The problem is there is no such thing as an NASB text that also has a 
> Hebrew Bible reference system. Since one does not exist, why are you 
> pressing for this ability? Mappings between reference systems should
be 
> handled in software systems and not markup.
> 
> I can almost see your last example, for Harry where Augustine has more

> than one historical reference system but in that case you just declare

> both editions and use two different osisRefs to indicated the
different 
> editions.
> 
> This is not a problem that really exists anywhere. If you want to do 
> mapping, do it at the software level and leave markup out of it. All
you 
> need for mapping is a single declaration of what text this "is" and if

> you have a map for that edition of that work, all is well.
> 
> Let me say it back to you: You want to have the ability to say that a 
> particular part of a text is known by different reference systems and
to 
> indicate in markup, the citation of that text in each of those
reference 
> systems. Is that fairly close?
> 
> So to use your example why can't I say:
> 
> <verse osisID="nasb:Ps.1.1><reference osisRef="nasb2:Ps.1.2"/>
> 
> Get the same results, unnecessary though I think they are, without
changing the present schema.
> 
> Patrick
> 
> 
> 
> 
> Todd Tillinghast wrote:
> 
>>In Toronto we determined that when resolving a reference that ALL of
the
>>elements in the <work> element in the referring document must be
present
>>in the THIS <work> element of the document being referred to.  
>>
>>REFERING DOCUMENT:
>><osisText>
>>	<header>
>>		<work osisWork="nasb">
>>			<title>NASB</title>
>>			<refSystem>Bible.Hebrew</refSystem>
>>		</work>
>>	</header>
>>	<div>
>>		<reference osisRefs="nasb:Ps.1.2"/>
>>	</div>
>></osisText>
>>
>>
>>TARGET DOCUMENT - CURRENT SCHEMA
>>This is a matching document because the THIS <work> has the only
>>required <work> element, "<title>NASB</title>", but multiple THIS
<work>
>>elements must be declared in order to provide for the multiple
reference
>>systems.
>>
>><osisText osisWorkID="nasbFirstCopy">
>>	<header>
>>		<work osisWork="nasbFirstCopy">
>>			<title>NASB</title>
>>			<creator>some text</creator>
>>			... more meta data ....
>>			<refSystem>Bible.NASB</refSystem>
>>		</work>
>>		<work osisWork="nasbSecondCopy">
>>			<title>NASB</title>
>>			<creator>some text</creator>
>>			... more meta data ....
>>			<refSystem>Bible.Hebrew</refSystem>
>>		</work>		
>>	</header>
>>	<div>
>>		<verse osisID="nasbFirstCopy:Ps.1.1
>>nasbSecondCopy:Ps.1.2
>>	</div>
>></osisText>
>>
>>Notice that the THIS <work> element must be stated twice but can only
be
>>indicated once (using osisWorkID in <osisText>).  The valid reference,
>>stated above, to this document can not be used due to the deficiency
of
>>the schema.  Both supported reference systems must be allowed to be
>>children of the SINGLE THIS <work> element.  
>>
>>
>>TARGET DOCUMENT - PROPOSED SCHEMA
>><osisText osisWorkID="nasb">
>>	<work>
>>		<title>NASB</title>
>>		<creator>some text</creator>
>>		... more meta data ....
>>		<refSystem id="nasb>Bible.NASB</refSystem>
>>		<refSystem id="Hebrew">Bible.Hebrew</refSystem>
>>	</work>	
>>	<header>
>>	</header>
>>	<div>
>>		<verse osisID="nasb:Ps.1.1 hebrew:Ps.1.2
>>	</div>
>></osisText>
>>
>>In this case the single <work> element that details THIS document has
>>more than one reference system.  Now when the reference identifies a
>>matching document based on the match of the needed child elements of
>>THIS <work>, the appropriate reference system can be selected from the
>>SINGLE <work> element that represents THIS work.
>>
>>Does this help clear things up?
>>
>>Todd
>>
>>-----Original Message-----
>>From: osis-core-admin@bibletechnologieswg.org
>>[mailto:osis-core-admin@bibletechnologieswg.org] On Behalf Of Patrick
>>Durusau
>>Sent: Wednesday, January 29, 2003 6:05 PM
>>To: osis-core@bibletechnologieswg.org
>>Subject: Re: [osis-core] RE: Work Issue
>>
>>Troy,
>>
>><snip>
>>
>>  
>>
>>>I'm not clear on exactly what problem we're trying to solve.
>>>
>>>No prefix means my default ref system (the refSystem inside my own 
>>>work entry), prefix means lookup the work in the header and use it's 
>>>refsystem, right?  This is true for both osisID's and osisRef's,
isn't
>>>    
>>>
>>
>>  
>>
>>>it?
>>>    
>>>
>>
>>That's what I think but cf. Todd's response and earlier posts. I am 
>>really trying to understand the problem Todd sees but quite honestly I

>>have not gotten there. Even in a long conversation in Toronto.
>>
>>Patrick
>>
>>  
>>
>>>    -Troy.
>>>
>>>
>>>_______________________________________________
>>>osis-core mailing list
>>>osis-core@bibletechnologieswg.org
>>>http://www.bibletechnologieswg.org/mailman/listinfo/osis-core
>>>    
>>>
>>
>>
>>  
>>
> 
> -- 
> Patrick Durusau
> Director of Research and Development
> Society of Biblical Literature
> pdurusau@emory.edu
> Co-Editor, ISO Reference Model for Topic Maps
> 
> 


_______________________________________________
osis-core mailing list
osis-core@bibletechnologieswg.org
http://www.bibletechnologieswg.org/mailman/listinfo/osis-core