[sword-devel] Legitimate FTP Mirrors & Module Distribution Rights Question

[Karl Kleinpaste]

"Troy A. Griffitts" <scribe at crosswire.org> writes:
> My apologies Karl.

None needed, Troy.

> I personally know that mobiles, swordweb, and BibleCS don't support
> maps (generally images) very well.

OK, well, that's not a critique of markup or format or anything else;
that's just features yet unimplemented or not yet implemented well
enough in certain UIs.  Gee, Xiphos still lacks av11n (due to be
resolved sometime next month, if my time doesn't get sucked away again
like the past 2 months).  PocketSword doesn't support genbooks yet.
Other UIs lack certain other features, I'm sure, though I don't keep a
tally.  Some day, I'd like to see at least one other image-comprehending
UI gain resize capability (and I don't mean just PS' pinch).

That said, PS (as a touchstone mobile app) does images just fine.

> We've discussed a better format than image modules for maps and some
> century might get around to implementing something, but Xiphos just
> goes out and releases something.  That's not a bad thing. I think you
> claim that characterization with your head high.

People who author modules for Xiphos repo generally want them out, now,
rather than when consensus is achieved.  Consensus, and feature
completeness, is never achieved in Sword.  So yeah, we put out stuff
more or less as soon as it's available.  If anyone's pet UI doesn't
grasp what such modules provide, I figure that's a problem for the UI,
not the modules, and it's an inducement to apply pressure for better,
more up-to-date implementation.  So many of the UIs have been feature-
stagnant for years.

I do test, both my own (quite few, these days) and others', especially
Brian's, because he does so much.  But the amount of testing to gain a
sense of "definitely worthwhile content, definitely looks to display
pretty well, no obvious problems in [say] internal cross-referencing, so
out the door it goes" is straightforward and not hard.

> Same goes with ThML vs. OSIS. Our policy is to work toward OSIS/TEI
> modules, but this takes ages. I'm happy for you to release before us
> in some other format, as it hasn't been a disruptions as all on our
> support list (likely due to the testing you mention). 

I know you want OSIS to gain the final upper hand.  And I know I don't
like using it.  We both have good and serious technical reasons for our
preferences, and we're just not going to achieve agreement on this, so
long as osis2mod's performance is routinely derided, and so long as the
OSIS spec itself is subject to such critique as could be found in Jan 2011:
http://www.crosswire.org/pipermail/sword-devel/2011-January/035683.html

Sorry, but I'm just not willing to abandon ThML's working and evidently
error-free capability in favor of an incompletely implemented standard
that isn't even self-consistent and whose authoring tools cause daily
grief to its users.  No way.  I know that OSIS provides conceptual
framework facilities that ThML inherently cannot, but until OSIS support
gets as good as ThML's, I won't touch it.  I have an extremely contrary
opinion that 99.44% of Sword app users are not particularly interested
in semantic markup but are deeply desirous of well-displayed content for
purpose of good Bible study.  I get that in ThML without having to drive
a truck through the OSIS swamp.

Now, with all that said, here's a way to get me to stop using ThML, that
I've mentioned now and again in the past:

Provide a ThML->OSIS converter.  A good one.  A reliable one.

> The final thing I didn't mention were modules like Gill. We've had
> trouble with tracking down a Gill module with a clean pedigree (well,
> maybe 'cleared' pedigree is a better word). 

Oh, yes.  Gill.

*sigh*

OK, look, Pierce knows it's out there, and he seems not to care.  If he
chooses to start caring, he's welcome to ask me about it.  I got my
original Gill from somewhere random (a formatting travesty -- a crime
against man, nature, and computer science) and just mechanically
converted it into something a human can actually stand to read.
Frankly, I don't see the issue in practical terms.

> Whereas, Xiphos is also braver is this regard to challenge legitimate
> copyright claims.

I'm not aware that any other module in Xiphos repo has even a remotely
questionable copyright status.  I have a bunch of personal-use modules
in a private repo that walk the fine line, but I don't generally
distribute those to anyone, other than an occasional developer who needs
a concrete clue or two on something that one of my modules breaks.

Anyhow...no big deal, Troy.  Confusion will be my epitaph.