Hey Ken,
Yeah, so, that's a question which requires a complicated answer. We've done research into a number of LXX 'editions' to determine book, chapter, and verse reference systems for each:
http://crosswire.org/svn/sword-tools/trunk/versification/lxx_v11ns/
The short answer is still: it's complicated. We are having to deal with this working on the Coptic Old Testament project, as well:
http://coptot.manuscriptroom.com/
Issues your project will need to consider. There are alternative traditions of the same book. Some editions labels these 'A' and 'B' and generally treat them as separate texts, with witnesses either witnessing to tradiation 'A' or tradition 'B'. Here is an example. Have a look at the CATSS edition and see the book names along the left:
http://www2.crosswire.org/study/bookdisplay.jsp?mod=LXXM
You can go this route and things work nicely in the VMRCRE for collating because the VMRCRE doesn't know that, e.g., JoshA and JoshB have anything to do with each other; witnesses are indexed as either JoshA or JoshB depending on their tradition. The downside is that often there are sections of these traditions which are shared-- common between the two, and this approach will not allow those common sections to be collated together across all witnesses of both traditions at these common points in the texts.
You can go a different route and define a superset of both traditions-- allowing max chapter count for the book and max verse count for each chapter to be the maximum of both traditions, and index all witnesses using the same book, e.g., Josh. This approach will collate witnesses from both traditions together. And while the VMRCRE generally does OK with this (especially variant graphs usually do a good job visualizing both tradition paths of text), it is usually not useful to see the same chapter and verse number from both tradition, e.g., Josh (tradition A) 18.21 usually has nothing textually to do with Josh (tradition B ) 18.21. In these cases, you could also run into some odd behavior with common words trying to be collated together across traditions when they have nothing to do with each other. It would be like collating Mark and Luke at Mark 1:1 and Luke 1:1. While you might find much shared text between the two Gospels, it is unlikely the shared text will be at the same points in the versification.
Finally, you can come up with a versification which chooses a (typically longer) tradition as the primary and index all witnesses using the primary versification for the longer tradition and all common parts of the shorter tradition, and then have a second versification for the uncommon parts, e.g., Josh, and JoshB-- Josh used to index the longer tradition witnesses and verses in common in the shorter tradition, and JoshB to index the remaining text portions which have no presence in the longer tradition. This gives you collation of all witnesses in the longer tradition and all shared portions in both traditions, and then gives you a second 'book' with only the remaining parts of the short tradition. The downside of this is that you get a second 'book' with only the remaining parts of the short tradition. This is a strange animal and isn't something which can be 'read' on its own. The shorter tradition needs to be pieced back together from the common parts plus the additional book to study the shorter tradition.
Anyway, it's all kindof a mess. You can talk with Felix and Frank at CoptOT to hear their current thoughts. Good luck! ;)
