On 26 Jul 2004 02:47:00 +0200, Erik Moeller <erik_moeller(a)gmx.de> wrote:
Yes, we should just all keep doing whatever we have
been doing for 20
years and be happy, and never try anything new. By the way, this whole
Wikipedia idea is kind of crazy. People should just use CVS to edit the
encyclopedia, why come up with something new when there is a perfectly
good process in place already? Jesus Fucking Christ, what kind of crap
attitude is this? "We don't have to guess or invent new ways"? No, we
don't, but guessing and experimenting is the source of all innovation.
As someone who insists on using basing logic in discussion, you set up
a nice strawman yourself.
You're right that the fact, that something hasn't been tried before
doesn't mean that it shouldn't be tried. If that was my only argument,
it would have been bogus.
I believe my arguments are a bit stronger than that. I claim that:
* there are better ways to improve mediawiki project
* we can find those ways by analysing what worked well for other
projects in the past
* bounty systems have been tried and failed, either miserably
(producing nothing) or just plain failed (dind't produce expected or
eve significant improvements, the word "significant" being the key)
Which you aren't even familiar with to the extent
that you could name
them. I can: CoSource, SourceExchange, the Free Software Bazaar. All these
models have been examined and have failed for different reasons. CoSource
failed for lack of VC, but had quite a lot of successful bounties
completed at the time it closed down (check
archive.org). SourceExchange
was targeted at corporations, not individuals. The Free Software Bazaar
was a static HTML website maintained by Axel Boldt, who lost interest at
some point, but it was reasonably active and led to some completed
projects.
I don't see how showing examples of failed projects that tried to use
bounty system supports your argument that bounty systems are a good
idea.
Whatever your opinion is on why they failed, the only objective
evidence we have is that they failed.
The open code market idea is certainly not new
I agree. I further claim that it has been tried and failed.
and it is one which is
increasingly being explored, refined and adopted, and which will
eventually inject the one thing into the open source process which is
currently missing from it: money. I advise you to at least do some basic
research and read, for example, Jordi Carrasco-Muñoz' paper "The Open-Code
Market":
http://firstmonday.org/issues/issue8_11/munoz/index.html
Thanks for the pointer. I've read the paper. It basically says:
"here's an idea and why I think it should work". Then I went to the
website (
http://www.opencodemarket.org/ which really is
http://jordiweb.net). As you can see by reading
http://jordiweb.net/docs/IT/The_Open_Code_Market.html (essentially the
same thing that was submitted to firstmonday) he published his idea on
May 2003. Over a year passed and nothing else has happened.
While I won't claim that his idea is worthless because he failed to
implement it in practice, I certainly don't see how you can present it
as a pro-bounty argument.
As well as "A history of markets for open source
software":
http://www.ms.lt/en/workingopenly/markets.html
Thanks, this is interesting information.
I knew about CoSource and SourceXchange - and they seem to me just
another failed dotCom business plans, a lot of hype, high-profile
launch and high-profile bust. Hardly an example of success.
Free Software Bazaar: "few dozen projects were completed, mostly for
about $50, one for $1,000". And they're gone. Again, hardly an example
of success.
Open Avenue - as far as I can tell total failure. No sign of even
modest accomplishments (not even completing a few dozen projects) and
the website redirects to an interesting (not work safe) place.
Asynchrony - the only thing that is alive today. But they weren't
really about bounty system, more like matching people with ideas but
unable to program with programmers, and taking a cut if they deliver
anything. Very safe business for Asynchrony but probably much less
successful for those who worked on projects. Regardless, today they
are just a IT service/consulting company. They changed the business
model.
Software Carpentry bounty to develop SC Config/Build/Test/Track. I
wouldn't call that a sucess either. Despite offering $200 k, the
system, to the best of my knowledge, has not been delivered.
They did pay some bounties for completed designs and
http://roundup.sourceforge.net is an implementation of SC Track
(although it's not founded by the bounty - it's an independent, unpaid
effort). So you might call it a partial success but as far as reaching
the initial objectives, it's just a failure. Additionally it (roundup)
shows that more was achieved by unpaid, motivated individual than by
the bounty system.
I don't
know a single succesful open-source project that implements
bounty system.
Your ignorance is unfortunate.
I agree and I thank you for providing links so that I could educate myself.
Now I can say that I'm familiar with several cases where bounty
systems were applied. And they all failed.
I'm still looking for an example of an open-source project that
successfuly implemented bounty scheme. By "success" I mean that a
significant portion of the code/functionality can be directly linked
to bounties offered. In particular, the GNOME and Mozilla bounties you
mentioned are, at best, failed experiments.
Both those projects failed.
GNOME bounty resulted in fixing 11 bugs
(
http://www.gnome.org/bounties/Winners.html). 32 bugs for which a
bounty was offered were *not* fixed
(
http://bugzilla.gnome.org/buglist.cgi?product=bounties&bug_status=UNCON…)
despite the fact that there were 5 months to do the work.
That's for a system that has around 700 new bugs opened weekly and as
much bugs fixed
(
http://bugzilla.gnome.org/reports/weekly-bug-summary.html) and is
approaching 150.000 bugs in total. Do you consider that a sucess?
Especially if you factor in all the work that went into nominating
bugs for bounties, creating web pages, evaluating the patches.
It also seems that it has been dropped by GNOME team - it was last
done in november 2003 and there's no activity aftter that. It's just
another failed experiment.
Mark has funded one project (
http://www.schooltool.org/) in
septemeber 2003. Calling it a bounty stretches a definition. If you
read the announcement
(
http://www.markshuttleworth.com/schooltoolbounty.txt) he's just
funding (almost) full-time programmers to work for him and develop a
software that is available as GPL.
The "real" bounties that are offered since at least december 2003 (see
e.g.
http://bugzilla.mozilla.org/show_bug.cgi?id=181866) for smaller
projects are still unclaimed.
Complete failure.
0 out of 5 Mozilla bounties were claimed. And Mozilla is progressing
just fine. Even if those 5 bounties were completed, the impact on
Mozilla would be insignificant.
Well, you can set up whatever straw man you want and
shoot it down, but
that has never been the point of the proposal. Read it.
The proposal is about offering bounties (or other rewards) for
completing developement tasks. As far as I know, this has not been
implemented *succesfully* in the past although many tried.
None of the examples you've presented can be called a success for any
reasonable definition of success.
What evidence there is that bounty system work *well* ?
Krzysztof Kowalczyk |
http://blog.kowalczyk.info