[osis-core] Trial encoding of TEV version of Matthew 13. Please read questions and issues.

[Steve DeRose]

At 01:16 AM -0500 06/05/02, Todd Tillinghast wrote:
>I have attempted encoding the TEV version of Matt.13 which has a fair amount
>of overlapping.  The sample encoded file and a modified schema are attached.
>Here are some issues. (NOTE: the attached schema is NOT an official release
>but is needed for the attached file to be valid
>
>Issue 1) How do we state the verseID of a partial verse?  Do we use the
>grain?
>Not that I used grain.  Comments?  Is style required?

I had been assuming it would have the same verseID on all parts, 
since they should all be retrieved when the verse is sought.

>
>
>Issue 2) We have accounted for overlapping hierarchies between verse and
>other elements as long as the VERSE gets split when there is a conflict.  If
>the verse element is kept UNSPLIT, then elements like paragraph need to be
>split and paragraph would need to be a child of verse.  This option seems
>ugly but will likely be preferred to the BCV oriented users.  (Same thing
>would be true for an other elements that a verse spans including div!  Do we
>want div as a child of verse?  If the answer is NO then it would seem that
>we should say that verse always gets split. Which is not complimentary to
>the BCV orientation.)
>
>Must the verse element always be split when there are conflicting
>hierarchies?

I would think so; the verses are the more arbitrary structural units, 
so just as they did when they were milestones, they would give way.

>Are there similar cases for <div> elements as chapters?
>
>
>Issue 3) Overlapping hierarchies that do not involve a BCV hierarchy
>conflict.
>Sub Issue 3a) Block quote overlapping with both div and paragraph.
>Option 1: Break up the Div and Paragraph at the start of the BlockQuote and
>have only one break.
>Option 2: Break the BlockQuote up into Paragraph size pieces.  (This makes
>the most sense to  me.)
>Trouble with both options:  What do we put in the next/prev?

2 seems better to me, too. Next and prev can be any value one likes 
-- they only serve to tie together the pieces.

>
>Sub Issue 3b) LineGroup, BlockQuote, Line, and Verse.
>Break Verse Up Strategy: Break up the verse into line size pieces and the
>hierarchy works  nicely.
>Maintain Verse integrity strategy: Break the LineGroup into two verse driven
>pieces.  (This  seems a little uglier.) (See prev post)

Per above, I favor breaking the verse.

>
>Issue 4) What value should be in next/prev for elements that are not verse?
>(BlockQuote, p, lineGroup, line, div, etc...)


anything.... I'd probably tend to make up an ID by using the element 
type as a prefix, and a serial number as suffix. Or maybe pack the 
verseId of the nearest verse in there too somewhere, so it's easier 
to avoid name conflicts without having to look around.

>
>(I created a block illustration of the various elements/hierarchies in
>Matthew 13 in Visio. If anyone is interested I could send it or create a
>JPEG version.)
>
>Todd
>
>Attachment converted: Sylph:Matt13NewB.xml (TEXT/ttxt) (000392C6)
>Attachment converted: Sylph:osisCore_1test10a.xsd (TEXT/ttxt) (000392C7)


-- 

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