[sword-devel] USFM conformance in usfm2osis.py

[Peter von Kaehne]

On 01/08/12 09:11, David Haslam wrote:

> There are some empirical aspects of USFM that are not specified in the *USFM
> User Reference*.
> They seemed to be defined /de facto/ by UBS Paratext.

I think we need to distinguish between accepted practice and simply poor
functionality in Paratext. Paratext is the reference implementation, but
lack of error warning in Paratext does not mean we need to implement
this as "allowed.

I would count \v 10-12 and, at a stretch, also \v 5,6 as accepted
practice - irrespective of what the spec says.

Similarly the \v 10a ...\s.... \v10b construct.

A fair number of other examples though seem to be essentially errors and
all parties benefit from us finding and correcting these. Bear in mind
that poor USFM leads also often to poor typesetting. So, our feedback
and corrections do help to fix the text.

Just because Paratext does not complain about a particular construct
does not make it accepted practice.

Peter