[sword-devel] Some questions about the WEB module

[Chris Little]

On 4/11/2012 5:12 PM, John Zaitseff wrote:
> Hi, Chris,
>
> Thanks for replying so quickly!
>
>>> 1. Who is the maintainer of this module?
>>
>> I maintain the WEB and associated translations modules.
>
> Should there be a wiki page somewhere to capture this information?  Or
> perhaps a simple line in the relevant web.conf?

I've never been partial to signing my work. Some people do, but most of 
the current bulk contributors of content do not. I'm certainly content 
that issues be reported through the usual support channels (JIRA and/or 
the support list), rather than being directed to individuals.

>> Do you know there to be meaningful updates to the text in the most
>> recent revision(s)--meaningful meaning not changes to spaces or minor
>> markup and not changes to the Apocrypha (since we don't support them in
>> the current WEB module anyway)?
>
> No, I don't know; I only have an old version from 2010, and I know there are
> changes from that.

It sounds like there are frequent updates currently. I'd prefer to limit 
updates for the WEB and related translations to one per month or so, so 
I'll do another update soon, but probably not for every change (at least 
until it's automated).

There should be a more recent version of the WEB module in CrossWire's 
main repository from March (or February, if not March), relative to what 
you have.

>> I use usfm2osis.pl.
>
> I presume with manual tweaking, as described in
> http://www.crosswire.org/wiki/Converting_SFM_Bibles_to_OSIS ?  If so, could
> you possibly send me your most recent OSIS file?

I don't do manual tweaking, unless there are errors in the source file. 
In that case, I fix the source file to generate valid OSIS.

I don't recall whether the currently posted WEB employs 
title-cleanup.pl, but I likely will not use it in the future. It's only 
needed if you're using recent builds of osis2mod, but I find those 
builds so poor that I just won't use them any more.

That's all to say that I don't intend to make any changes to the output 
of usfm2osis.pl, but I'll send you the file in the event that I actually do.

> A related question, by the way: how well does mod2osis work when the
> original was an OSIS file?  That is, if file.osis is processed thus:
>
>    file.osis ->  osis2mod ->  mod2osis ->  newfile.osis
>
> How close is newfile.osis to the original file.osis?

Not at all. mod2osis is worthless.

>> One of my personal Summer projects will be to write a USFM to OSIS
>> converter from scratch, but for that I won't be touching the
>> usfm2osis.pl codebase.
>
> This was precisely what I was talking about.  However, summer is a long time
> away, especially down under in Australia! :-)  A few questions:
>
> 1. What is technically wrong with the script at osis-converters
>     (http://code.google.com/p/osis-converters/)?
>
> 2. If there ARE technical limitations with sfm2osis.pl script there, is it
>     better to work with that than start over yet again?  How amenable is that
>     maintainer to accepting patches?
>
> Beware of the NIH (Not Invented Here) syndrome!  I occasionally fall into
> the trap myself, but I _am_ trying here to work with people.

Summer comes after school for me, which is 2-4 weeks from now, depending 
on how much studying I need to do for finals and how long grading papers 
takes.

I've never looked at the osis-converters scripts, so I can't comment on 
them at all. However, it wouldn't be appropriate to my goals, since I 
specifically intend the rewrite to be in Python and my intention is to 
convert USFM to OSIS via a USX. So my real process will be:

USFM -> USX (assuming USX isn't the source format)
validate USX
USX -> OSIS
validate OSIS
plus sanity tests (checking for duplicate/missing verses, balanced 
sID/eID pairs, etc.)

--Chris