[jsword-devel] Vulgate v11n mapping

[DM Smith]

I think I've found some errors in the o.c.j.versification.Vulg.properties file.

I'm ignoring the verse 0 mapping bug in the following discussion.

I'd like verification. I'm attaching the modified file.

I went through the file entry by entry, comparing CPDV to the KJV modules. The advantage of CPDV is that it is English and I can readily validate the versification. I'm making the assumption the CPDV is representative of the Vulgate v11n.

Psalm 2
There are no mappings. But it appears that
	Vulg.Ps 2:12-13  = KJV.Ps.2.12
Without this mapping I'm getting an NPE. Apparently the default mapping of Vulg.Ps.2.13 = KJV.Ps.2.13 is causing this.

Psalm 3
Vulg.Ps.3.9 is not present in the KJV. I think this should be:
Ps.3.9=?Missing
When I have this, I'm getting errors that prevent the chapter from showing.
Having:
Ps.3.9=Ps.3.8
works (mapping both Vulg 3.8 and 3.9 to KJV 3.8) but I don't think this is right.

Psalm 4
This has a canonical title in the KJV, but there is no mapping for this. I think the following should be added:
Ps.4.1-Ps.4.8=Ps.4.0-Ps.4.7
With this the verses line up as in others that have a canonical title.

Psalm 28
This has the following mapping:
Ps.28.1=Ps.29.0-Ps.29.0
This is preventing KJV verse 1 from showing. Obviously the second .0 should be .1
Ps.28.1=Ps.29.0-Ps.29.1

Psalm 98
This maps to KJV.Ps.99 but there is no mapping in the file. The impact is that it uses the default mapping to KJV.Ps.98.
This should be:
Ps.98.1=Ps.99.0-Ps.99.1
Ps.98.2-Ps.98.9=Ps.99.2-Ps.99.9

Psalm 99
This is rather complex. The Vulgate verse 1 is the KJV canonical title which is pre-verse in verse 1. Vulgate verse 2 is KJV verse 1 *and* verse 2. 
However it is mapped as:
Ps.99.1=Ps.100.0-Ps.100.1
Ps.99.2-Ps.99.5=Ps.100.2-Ps.100.5
Which treats Vulgate verse 1 as the merge of KJV's canonical title and verse content.
This should be something like:
Ps.99.1=Ps.100.0
Ps.99.2=Ps.100.1
Ps.99.2-Ps.99.5=Ps.100.1-Ps.100.5

Psalm 108
This is just like Psalm 99.
However it is mapped as:
Ps.108.1=Ps.109.0-Ps.109.1
Ps.108.2-Ps.108.31=Ps.109.2-Ps.109.31
It should be something like:
Ps.108.1=Ps.109.0
Ps.108.2=Ps.109.1
Ps.108.2-Ps.108.31=Ps.109.2-Ps.109.31

Song of Solomon 5-7
This is complex. The Vulgate and KJV have different chapter boundaries.
The last verse of Vulgate.Song.5 is the first verse of KJV.Song.6. This is correctly represented.
The last verse of KJV.Song.6 is split between the last verse of Vulgate.Song.6 and the first verse of Song.7.
Then Vulgate verse 2-ff should map to KJV verse 1-ff.
Additionally, KJV verse 12 is not part of the Vulgate.
It is correct that verse 13 is an identity mapping.
The mapping has:
Song.7.1=Song.6.13!b
Song.7.1=Song.7.1
This incorrectly maps KVJ verse 1 to Vulgate verse 1.
This should be:
Song.7.1=Song.6.13!b
Song.7.2-Song.7.12=Song.7.1-Song.7.11
There is a problem with KJV.Song.7.12 It doesn't show with KJV first or second in parallel view w CPVD.
There should be an entry like:
?=Song.7.12
But that doesn't make it show up, either.

Daniel 3
Deuterocanonical material is in the middle of the chapter in the Vulgate. Essentially, extra verses are from 3.24-3.90.
However, there is no mapping for this deuterocanonical range. The impact is that KJV content shows up in parallel view for 3.24-3.90, going through chapter 4 and into chapter 5.

There should be an entry like:
Dan.3.24-Dan.3.90=?deuterocanonical
But this doesn't work. Causes errors.

I don't know if any of this was from changes that I made to the mapping code. (Such as having ? on either side of =.)

Some of the things I have learned from this exercise:
o It is really easy to make mistakes.
o If identity mapping is not true, there needs to be a mapping for every verse in both v11ns. If the chapters don't have the same number of verses then merge, split or missing has to be in the file.

In view of the last one, I think a good test is to check the mapping to ensure that every verse in a v11n is covered by the mapping both forward and back. That is if a verse is mapped, then both from and to should be valid. If a verse from the v11n is not in the map then the identity must exist in the other v11n.

Comment on the verse 0 bug. Currently the mapping files use verse 0 to represent the pre-verse content of verse 1. This hides the SWORD/JSword meaning that verse 0 means book introduction (with chapter 0) or chapter introduction. The simple fix is to add an explicit verse 0 mapping. The better solution, that we've agreed on, is to change the mapping to use Book.n.1!pv instead of Book.n.0 to reference the preverse and Book.n.1!v to represent verse content. Then 0 means 0.

In Him,
	DM

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.crosswire.org/pipermail/jsword-devel/attachments/20140402/167a0bd0/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Vulg.properties
Type: application/octet-stream
Size: 7334 bytes
Desc: not available
URL: <http://www.crosswire.org/pipermail/jsword-devel/attachments/20140402/167a0bd0/attachment.obj>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.crosswire.org/pipermail/jsword-devel/attachments/20140402/167a0bd0/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 4145 bytes
Desc: not available
URL: <http://www.crosswire.org/pipermail/jsword-devel/attachments/20140402/167a0bd0/attachment.p7s>