[sword-devel] Getting stuff done (Re: External links)
karl at kleinpaste.org
Wed Nov 26 07:14:43 MST 2008
Spare us the "they're not *really* ThML" whining. I don't care that
they're specifically ThML. When I first started producing Sword modules
2 years ago, I was using GBF.
The only important distinctions about my modules are:
- They work.
- They get out readily to people who want them.
- They provide utility in their Sword apps.
- They give them understanding and knowledge that they didn't have last
year or last month or last week.
My repo dumps out ~360Mbyte/day (high std.dev.), or ~10Gbyte/month.
No doubt the Crosswire repo ships a great deal more than that, but
considering that my repo gets no visibility other than the Module
Repositories wiki page and word-of-mouth advertising, I believe that
says a great deal.
I DO NOT CARE if they meet some standard of perfection. IT DOES NOT
MATTER if they validate. IT IS OF NO RELEVANCE that they're not clean.
What matters is that I, and we of GnomeSword development, understand the
words "DONE" and "WORKING" and "USABLE" and "RELEASED" and The Sword
Project as a whole does not.
As SourceForge says in self-descriptive text, the mantra of open source
software is "release early, release often." The Sword Project mantra
could nearly be copied from The X-Files: "Release never, because The
Perfection Asymptote Is Out There." As just recently said, The Sword
Project can "take time to make sound decisions," which is a lovely
thought, except that the actual effect, the actual practice observed is
that decisions are seldom actually made, as demonstrated by the fact
that results are too infrequently seen in released code. That's why
much weaker programs like e-Sword are wildly popular (v.8 just released,
touting i18n as a new feature) and The Sword Project, though older, more
mature, and more capable in concept and structure, can't get anyone's
attention even when we jump and down in front of their faces.
> So write a patch for some filters.
You want better OSIS filters, you write them. I don't need them.
I have no time -- too occupied with GnomeSword as it is. I did an
interesting exercise the other day: "svn blame" on the entire content of
GnomeSword. I found that, since late 2006, I have written 20,000 lines
of code. I have been writing ~740 lines of code per month on a project
that doesn't even pay my bills. That doesn't count the time I put into
module production: I have 75 modules in 2 repos, only half of which I
can give away -- I produce a new module on average roughly every 2-3
weeks. They don't get put through any painful beta period, they just
get produced and put up for use. People tell me about problems, I fix
them, a new version goes up with an updated stamp, and life goes on.
And I get nice thank-you notes from people for them.
I'm scratching a deep personal itch in large proportion. My itch is
most emphatically not "perfect tools in perfect syntax for perfect
analysis by perfect academics, to be released only and exactly on an
unknowable and impossible perfect future day when all actually achieve
My itch is "more people doing Bible study using good tools and quality
BibleCS is not a good tool. No reviewer likes it -- not one. The
Foster review (which keeps getting updated, he's added a couple in just
the last week) originally put BibleCS at 10th out of 12 IIRC. Because
of additions since original discussion in April, BibleCS is now 13th out
of 17 -- 5 more packages reviewed, 3 slots farther down in the list.
Significantly, the Foster review concluded with this observation: "The
interface is poor, and I personally do not think this is a version 1
product, let alone 1.5.9...I cannot recommend this product at this time,
there are simply many other superior and free packages available." That
speaks not mere volumes, it speaks whole encyclopedias.
My itch is for GnomeSword-WIN32 to be DONE and WORKING and USABLE and
RELEASED by Christmas. We have our own plans, our own timing, and our
own sense of our obligations, which have nothing whatever to do with
whether The Sword Project as a whole will produce another release of the
engine, whether filters will be written, whether argumentation continues
over how TEI might be subsumed into the greater OSIS, or anything else
of such a sort.
All that matters is that we see a finish line and we're bound and
determined to cross it.
We Get Stuff Done.
More information about the sword-devel