[sword-devel] modules upload etc - suggestion

[Karl Kleinpaste]

I think Peter's suggestions are very good, for the most part.

The lack of an "alpha" area where authors can make more or less
immediate updates when problems are addressed is a serious sore point.
It's why I put up a public repository of my own in the first place,
where a half dozen of the 30 modules available are not built by me --
because things go into today's beta repository and never come out, and
once there... well, everyone reading this list has seen me ask for
certain updates to be done, and has seen how long it takes any of those
requests to be handled.  Case in point, some 3 months ago, discussion
was heard about whether the GnomeSword manual module actually belongs in
the Crosswire repositories at all, it was generally decided that
ultimately that's not the right place for such a thing, and I (lacking
privs) asked that someone just remove it.  It's still there.  And aging
steadily: 2.3.2 is in my own repo (minimal updates), and 2.3.3 will
happen this Sunday (quite a few updates).

By comparison to the beta repository, Peter could give me an updated
FarBibAtlas or FarTV this evening and it would be available to the world
by sometime tomorrow.  But wouldn't it be best if Peter could give the
world his update on his own?

This past Sunday, I wrote some paragraphs to Troy privately, to talk a
bit about an automatic upload facility that could be integrated into the
Sword lib without really much difficulty.  (I may try my hand at that
code next month, but at the moment I'm cramming a lot of work into
getting structural GS improvements made before BibleTech.)  And if that
were in place, my GS changes to put it to use would be 20 or 30 lines of
code, and possibly not that much.  As a technical problem, this
"publish" concept is not too tough and it just leaves the political
problems, such as DM mentioned -- notably, malicious people uploading
crap, or uploading modules that can't properly be distributed by
Crosswire.  That's largely solvable in a technical manner, too, by
careful setting of the upload repository's permissions.

Assuming that perms on the upload repository are clamped down, the
remaining technical/political/personal combination problem regards
putting in some small bits of infrastructure so as to notify the
repository owner of newly-arrived modules in need of being approved and
pushed around appropriately.  A quick first pass at that resulted in an
adequate if not very featureful cron-driven shell script that does the
job -- less than 30 lines.  Indeed, if there existed such an upload
capability in the Sword lib today, I would be maintaining an open
"publishers' repository" by this weekend, directly usable by at least GS
and any other UI that implements the interface.  (BTW, if one cares to
leave the perms rather wide open, then the upload repository is also a
download repository, for everyone else to get at the new modules.)

I am thinking through these sorts of problems because I am again
considering how to integrate the idea of users authoring their own
modules, and how they could get them published in some sense, where
friends/students/the world could get at them.  Whenever I have queried
GS users about needed features, authoring tools have consistently been a
top request, and I have been woefully lax about getting to it.  There
are 150 Bibles available, but it would be nice if any respected author
could distribute a sermon series or Bible study series in Sword format,
in commentary or genbook form.  It would enhance Sword's PR value
immensely if people saw that they could readily contribute to the body
of work available for study.  I believe that as long as it's difficult
for Joe Random to produce and distribute a Sword module (cf. Peter, who
succeeds because he is bull-headed enough to keep pushing), that kind of
community critical mass will not come into being.