Re: LoMuS

[Graham Percival]

On Fri, Apr 27, 2012 at 07:09:39AM +0200, address@hidden wrote:
>    I'd be good to set a precedent for this now so that LilyPond can apply
>    to other software competitions in the future.

*sigh*
I'll rehash an old email.


I know that some of us would like to get paid for lilypond work.
But I'm leery of jumping into this topic for a few reasons:

1) lilypond survives because of volunteer work.  I think we should
be *extremely* cautious about anything which may jeopardize the
amount of time people volunteer.

2) if we officially organize these arrangements, there may be
currency exchange, tax issues, or legal issues that arise,
particularly given the international nature.  Who gets the money?
How would it be distributed, if at all?  How much loss to third
parties is acceptable?

3) it's not going to be enough.  Companies and government grants
are where the real money for supporting programmers comes from;
hiring just one person from individual donations would require far
more users than we have.


1) VOLUNTEER EFFORT

I'm referring to "maintenance" tasks here, not "new development".
I famously spend 10 hours a week[*] just keeping things running.
Add in the bug squad, patch countdowns, building releases,
reviewing patches, miscellaneous things that James does, etc, and
we're easily looking at 30 hours a week just keeping things
moving.

[*] now 5 hours due to phd dissertation.

That may sound like a lot -- well, it *is* a lot -- but other than
automation like Patchy, I can't see how to reduce it.  We could
ignore all bug reports, abandon any pretense at avoiding
regressions, stop having releases, eliminate any patch reviews...
at various points in the history of lilypond, we've done all of
those... but I really think we should keep those tasks going.

How much does that cost?  Well, if we wanted to pay people $10 an
hour for those tasks -- which is probably decent for bug squad
work, but ridiculously low rate for reviewing patches -- then
we're looking at $1200 a month.  For *no* new features or bug
fixes.  A thousand bucks a month just to keep things moving
smoothly.

We can't afford to jeopardize that.  We need volunteers willing to
handle those tasks.


Some people aren't going to make a priority of seeking out
monetary gain; they have good careers already.  Other people might
have crappy jobs or be students and would probably love to have a
bit of extra financial benefit.  I'm not certain how those people
might feel about spending hours each week doing volunteer
maintenance tasks if programmers were working for money.

Personally, at this stage in my life, I'm more in the first group
-- I'm (on track) for a good academic job, so I'm not too fussed
about money.  Also, I'm single and live on less than 500 pounds
($700) a month.  But if a large fraction of development work was
on a "for hire" basis eight years ago, I would not have gotten
involved in lilypond development.  So I think that concerns about
jeopardizing volunteer effort with financial benefits for some
people are very real.



2) OFFICIAL PROBLEMS

If the lilypond project officially organizes/endorses
sponsorships, are there any legal issues?  I _think_ that as long
as we treat programmers as contractors, the programmers themselves
are responsible for filing taxes with their own governments... but
I'm not a tax lawyer (much less an international tax lawyer!) so
I'm not certain.  What currencies would we accept, and which
currencies would we pay out?  Would we go with a specific tool
(google checkout?  paypal, despite any ethical qualms people might
have with that company?  etc) ?  organize bank transfers within
Europe (which are much easier than bank transfers within Canada or
the US, and certainly easier than international bank transfers).

This would open up a *huge* can of worms, and I don't want to deal
with any of this crap.  There's always the possibily of making /
resurrecting the lilypond foundation (yet more crappy paperwork),
or joining a group like the Software Freedom Conservancy.

The SFC actually looks decent -- I recognize a bunch of names on
their directors, and some of their projects are boost, git,
inkscape, pypy, swig, and wine.  But that would still involve lots
of discussion, paperwork, and red tape.

I know that some of you might think that's exactly what GOP is --
and yes, I've been thinking about proposing this in GOP 25 or so
-- but right now I don't think it's a problem worth tackling.  We
have enough problems with staging / not losing patches / etc.
Once we've cleared out those and have a smoother development
process, I'm open to investigating the SFC more.

... ha!  I mean, the whole point of GOP was to get the development
process moving smoothly enough so that I could turn my attention
to GLISS.  I really, really, *really* don't want to mess around
with official foundations at the moemnt.


3) NOT ENOUGH

The only source of enough money to really make a difference -- to
hire programmers on a part-time basis, let alone full-time -- is
commercial support or government grants.

FreeBSD is quite successful at getting commercial sponsorships;
many web-hosting companies use freebsd.  I'm not aware of any
large companies using lilypond, so that option is out.

I think it's quite realistic to get serious money from artistic
grants.  There was that Russian programmer who spent 6 or 12
months working on lilypond for some folk music musicologists...
but unfortunately he didn't put much effort towards sending his
patches upstream.  I could imagine either teaming up with similar
musicologists (particularly "niche" markets, such as chant, folk
music, etc), or teaming up with contemporary composers.  In some
ways, you could consider Mike to be doing this already, since he
programs stuff to support his compositions, which I think are
indirectly paid by the government (via schools or competitions?).


However, applying for grants, or at the very least, making
contacts with academics who are applying for grants, requires more
paperwork and/or networking.  Don't look at me for that.

As far as sustained cash from our existing user base of
individuals... I think we could reasonably expect about a thousand
a month.  Maybe as much as two thousand.  Which might sound like a
lot to a student, but $24,000 a year for an experienced software
developer is nothing.



I really wish that we kept things at the individual level.
Developer offers his services; some users send him money.  Some
people apply for a music history grant to archive some chants and
improve software to do so; they get money and work on the neume
code in lilypond.  etc.  All private deals.

If you want to have lilypond involved at the organizational level,
we need an actual organization.  Setting up an official body for
lilypond is certainly possible.  It'll take roughly 200 hours.
We'd need to pick a country, study that country's laws on
non-profit (or profit?) organizations, probably create a board of
governers, fill in some paperwork, pay some fees, probably have
meeetings with an agenda and stuff, find out if it's possible to
hold such meetings via irc or skype and maybe pick a different
country based on the answers, hopefully pick a country that has
laws in English or else rely on google translate for legal
documents, etc.

On second thought, the estimate of 200 hours is probably too low.

Even if we had a legal organization, we'd need to figure out how
the finances would work.  Who donates our domain name registration
and bandwidth?  Would they care if we got money?  Maybe we should
look at commercial hosting instead,  How much money does our
bandwidth cost?


And then of course we come back to volunteers.  If there was a
formal organization with income from grants and users and the
like, how would that affect the people currently volunteering
their time?  To me, this is the most important point.  Maybe we
should run an anonymous survey to find out how many people would
reduce or eliminate their volunteer work with lilypond?  The
answer is "at least one".

Pushing this financial stuff is raising the stakes.  The gamble
might pay off greatly, or it might cost us a lot.  I don't think
we have enough information to make any reasonable estimate on what
the odds are.

- Graham