[osis-core] [Fwd: OSIS Schema Issues]

[Patrick Durusau]

  Guys,

Forwarding a list of outstanding issues I summarized for Steve. We are 
talking later today and will report/discuss on the conference call 
tomorrow as well.

Corrections or additions (be brief) welcome!

Hope everyone is having a great day!

Patrick

-------- Original Message --------
Subject: OSIS Schema Issues
Date: Thu, 05 Jun 2003 09:34:30 -0400
From: Patrick Durusau <pdurusau@emory.edu>
Reply-To: Patrick.Durusau@sbl-site.org
To: Steve DeRose <sderose@acm.org>



Steve,

The outstanding issues I have culled from posts since the Dallas meeting:

1. Milestone syntax (elements that are normally containers but written 
as empty elements)

Stock text: In the beginning...

Current schema: In the beginning...

Dallas proposal: In the 
beginning...

Todd's proposal: In the 
beginning...

Todd argues that he needs to know whether the milestone he has found is 
a beginning or ending milestone. I think this is trying to answer the 
fragmentation issue, which is actually broader than simply the use of 
milestones.

Troy agrees that we can create an atributeGroup but would prefer:

// stole regex from osisWorkType

  
    
  



// all documentation about Milestoneable behaviour here

  



// example element

  
    
      
...
    
  


Does have the advantage of making it clearer which elements can be used 
as milestones. We have not used this mechanism elsewhere (not an 
argument against, just an observation).

2. scope attribute on div?

Todd wants to have a scope attribute on 
that uses the same regex as osisRef to indicate the range of references 
in a
. Chris agrees but wants to add that scope should allow a list of 
osisRefTypes. 3. osisRef required on ? Todd suggestes that osisRef be 
required on . Chris opposes as being optional allows marking references 
to work that have no reference system. 4. as child of

Troy gave the best statement of the problem: ***Troy*** I do understand 
your confusion. But to help clarify the problem. The BCV numbering from 
centuries ago sucked. MANY chapters begin and end in ridiculous places. 
Modern translations add sections and paragraphs where the translators 
feel they are logical (which still varies among translator teams). So, 
ANY TIME they disagree with the traditional BCV divisions of 
thought/subject changes (which is quite often) there will be paragraphs 
that start mid 'chapter' AND 'chapters' that span paragraphs. They are 2 
different hierarchies. The first is modern translation decision, the 
second is traditional BCV numbering. ***/Troy*** Todd has suggested the 
following solution: ***Todd*** I think the bottom line is that 
milestones can occur just about anywhere and we need to support the 
ability for them to exist just about anywhere. At the same time we are 
not comfortable having a container just about anywhere and in fact that 
is not how we would intend the container to be used. One possibility is 
to create two Complex Types for chapter. 1) One that is like

and would be generally used and 2) One that has no children and is 
always empty for use as a milestone everywhere is allowed. The encoding 
using as a milestone would work like it needs to but we would not be 
tending toward an ANY model, while at the same time preserving 
consistent markup representation of the multiple overlapping 
hierarchies. The following "best practice" encoding

...abc......def...

...hij...

could be viewed equivalently by software that provides multiple DOM 
views as

...abc... ...def...

...hij...

This alternant encoding does not include any "unusually" children of as 
a container AND the first encoding does not have a milestone in an 
unusual location. CONCLUSION: By creating a special "milestone only" 
that is a possible child of

, , , , ..., we can truly support the two hierarchies that can overlap 
at just about any point, while not opening the model up to an ANY type 
encoding. ***/Todd*** I am uneasy about Todd's solution because I 
thought we agreed that B/C/V was the traditional and gave way to B/S/P. 
What I think is happening is that Troy/Chris are using 'chapter" to mean 
section in the modern sense. I am hearing "chapter" in the traditional 
sense. Not sure how we can educate people that "chapter" means the 
traditional division and that they should not use it for modern 
translation breaks into "chapters" which are really sections to which 
the traditional divisions give way. If we really are confusing "chapter" 
then the solution is to require B/S/P and do milestones for the 
traditional B/C/V. 5. Local reference extension: Do we want to use ! or 
| and can you remind me of the syntax that I need to add to the regex? 
Any other issues that we need to settle for the schema itself? (As 
opposed to best practices?) Hope you are having a great day! Patrick -- 
Patrick Durusau Director of Research and Development Society of Biblical 
Literature Patrick.Durusau@sbl-site.org Co-Editor, ISO 13250, Topic Maps 
-- Reference Model Topic Maps: Human, not artificial, intelligence at work!


-- 
Patrick Durusau
Director of Research and Development
Society of Biblical Literature
Patrick.Durusau@sbl-site.org
Co-Editor, ISO 13250, Topic Maps -- Reference Model

Topic Maps: Human, not artificial, intelligence at work!