[sword-devel] osis2mod output (Bisaya-Inunhan NT)

[DM Smith]

On Apr 29, 2009, at 7:57 PM, Greg Hellings wrote:

> On Wed, Apr 29, 2009 at 6:45 PM, Chris Little  
> <chrislit at crosswire.org> wrote:
>> Jonathan Marsden wrote:
>>> Similarly, since one can specify a versification system inside the  
>>> OSIS
>>> XML file, why does osis2mod need a -v switch to select a  
>>> versification
>>> system?  Can't it just use the one specified in the XML input (and  
>>> default
>>> to the OSIS default, if one is not specified in the OSIS XML)?
>>
>> There's no guarantee that 3rd parties will identify reference  
>> systems in the
>> same way as we do. There's no requirement from OSIS that reference  
>> systems
>> conform to some standard set. And we don't want to prevent people  
>> importing
>> data that does not conform to one of our explicitly supported v11n  
>> systems.
>>
>> If DM wants to add the capability to osis2mod to catch the ref system
>> declaration and understand Bible.(KJV|Leningrad|Tanakh|NRSVA|...),  
>> that's
>> fine, but we definitely need to allow explicit overrides via a  
>> switch as we
>> currently have.
>
> Not sure I agree with this.  If the OSIS file uses Tanakh and claims
> that is uses GNT, then even if the user wants to change that on the
> command line, they shouldn't be able to.  To do so would clearly
> violate anyone's (within reason) "best practices" and we are trying to
> encourage and help people to use "best practices" when encoding their
> OSIS.
>
> Certainly, having a manual switch in case the OSIS file doesn't
> specify might be good, but even that would not be "best practice."
> The "best" thing to do would be follow the OSIS standard if that's
> what we want people to use -- assume NRSVA unless the file states
> otherwise.  Anything else is clearly not the "best" in regards to the
> semantics of the specified OSIS standard.

All good input.

Many thanks.

I know I won't remove the -v flag. It's use should override any  
statement that is in the OSIS file. As suggested if it differs from  
what is in the header, a message should be appropriate, especially if  
what's in the header is a supported versification.

My guess is that people will cut and paste a header w/o much thought  
to anything but getting the OSIS xml file to validate. Maybe I'm too  
cynical.

 From what I gather the NSRVA is the OSIS standard as a single point  
of mapping. As such it is not merely the NSRVA, but a one with  
additions, so that it can be that central point of mapping.

So while the NSRVA could be the default to osis2mod, I think it would  
induce laziness, when module makers should find the closest v11n match  
to their input.

I'm not up to it, at least in the near future, but someone should  
create a tool that will analyze an OSIS input file and suggest the  
best match.

Keep the comments coming. It only makes us better as we work together.

Working together in Christ to advance His Word,
	DM