[sword-devel] OSIS and intros

pinoaffe pinoaffe at gmail.com
Fri Jul 11 16:33:50 EDT 2025


DM Smith <dmsmith at crosswire.org> writes:
> The issue with heuristics is that they essentially enforce
> assumptions, which may very well be wrong.
I think a heuristic along the line of "Does this book group contain only
new testament books and are all new testament books that are present in
this bible contained in this book group" might be good enough for our
purpose

> The module team has a usfm2osis.py to convert USFM to OSIS. Michael
> also has one. The nature of USFM is that it’s goal is presentation not
> OSIS’s nature of semantic markup. I don’t know how that transformation
> handles intro material. I’m pretty sure neither has a begin/end
> testament construct. It is a simple thing to manually add such after
> transformation if that fits with one’s workflow.
>
> I think it’s a shortcoming of the OSIS spec that it is flexible when
> marking up testaments. Your suggestion of explicitly marking testament
> may be the only real solution. In the KJV’s OSIS it uses
> type=“bookGroup” subType=“x-{testament}” where {testament} is OT or
> NT. It also uses x-DC for the apocrypha.
I think OSIS's flexibility is also its strength, particularly when it
comes to milestoning and giving module authors freedom on how to
structure their bibles - otherwise many bibles could not be represented
in OSIS.  However, an additional list of standards/conventions would be
good to add, at the very least something like "If you have a testament
bookGroup, represent it like this, if you have a gospels bookgroup,
represent it like this, etc.

In general, I think it would be good if the OSIS spec got a major
overhaul.

> I’m wondering whether SWORD’s [ Testament 1|2 Heading ] could/should
> be represented as an OSIS ID, type or subType.

I've not put much thought into this, but have no strong preference

> Should a mechanism be imposed on OSIS creators? We’ve resisted that as
> much as possible.
I feel like it would not be much of an imposition.  We could even add an
(optional?)  preprocessing script that tries to group books by testament


More information about the sword-devel mailing list