Conrad Irwin wrote:
On 03/31/2010 12:31 PM, Victor wrote:
Hi, now I see...
I've posted a message to the fa.caml newsgroup:
http://groups.google.com/group/fa.caml/browse_frm/thread/1593e053759d7679
hopefully somebody will volunteer to fix the issues, thus
saving the human resources for better tasks.
While I'll no doubt regret saying this, I am happy to fix some of these
bugs. With the majority of these it's harder to decide "should we fix"
and "how should we fix" rather than actually being hard to implement.
What I don't want to do is fix things to find that it then gets
immediately re-implemented in PHP, which seems to be what people want.
I don't think "reimplement texvc in PHP" is anyone's goal as such. The
real goal is "make texvc actively maintained". Reimplementing it in PHP
would be one way to achieve that, since we have plenty of PHP
programmers who could then maintain it. Finding some person or people
who know OCaml and are willing to do the work would be another route to
the same end.
In general, there's a form of decision paralysis common to volunteer
projects, particularly ones relying on skilled volunteers. Basically,
there are two solutions to a problem, X and Y:
Person A says "I can try to do X, but I don't want to spend the time if
we're just going to do Y instead."
Person B says "I can try to do Y, but I don't want to spend the time if
we're just going to do X instead."
No-one else wants to commit to either X or Y, since they want to keep
the other option open in case A or B doesn't succeed with their favored
approach after all.
End result is that neither X nor Y actually gets done.
There are two ways out of this situation: either the project needs to
commit to one option and make sure it gets done, or A and B need to
accept the risk that they might end up doing redundant work.
Ironically, the meta-decision on whether to commit to one approach or
try both can also get suck in a similar dilemma on a higher level.
Going back to the concrete issue here, I'd personally recommend trying
both _for now_. In particular, rewriting texvc in Python or PHP, as a
MediaWiki extension, would seem like good GSoC project even if it didn't
actually end up being adopted into MW core in the end.
Meanwhile, fixing at least the simplest and most critical issues in the
current OCaml implementation would also be of immediate value, even if
that implementation might possibly end up being replaced at some point
off in the future. I wouldn't necessarily recommend going immediately
for the more tricky issues, or the ones with lower short-term benefit
per effort, but I'm sure there must be some low-hanging fruit ready to
be picked by anyone who's simply familiar with the language.
By autumn, we ought to have some idea how much, if any, progress has
been made with each approach. At that point, we should be better able
to decide whether to commit to one approach or the other, and if so, which.
I should also note that, as long as one implementation doesn't
_completely_ supersede the other in every way, there would probably be
people interested in using each of them if they were available as
optional extensions. In particular, I'm sure there are people who have
access to Python but haven't managed to set up OCaml -- and I wouldn't
be completely surprised if the opposite turned out to be also true.
Of course, that's just my opinion as a random occasional contributor.
Take it with as much salt as you think appropriate.
--
Ilmari Karonen