Re: Funding GPL projects or funding the GPL?

[Federico Ferreres <fferreres@ojf.com>]

It's $10 or $20 per user per year to be able to use fGPL code _alone_.
In addition to that, you'll have to pay for any extras (services or non
fGPL code) you may like purchasing: the nice Red Hat Network services,
your Debian CD and my Slackware CD (but I would preffer to just fund
Patrick though fGPL instead). The membership just allows you into "the
club".

Also, publishing something under the fGPL doesn't mean the developer
will get the funds, unless people choose to fund it or if it's a core
app (this last decided by a board of directors for example, lead by
Linus or whoever is most respected and wants to get involved and up to a
fixed % of the yearly fund raising). 

This doesn't kill any other funding source you already have established.
IBM may not want to fGPL webalizer, Covalent or MySQL AB as well. And
that's fine. Copyright owners can offer them other licenses suitable for
them (X style, BSD, whatever). Also if IBM (or any other company) wants
to fund a different project because it makes sense for them, and users
and small companies do not want to pay for it, then they will fund it
themselves, as it is the case today.

Now some figures: if you have 50.000.000 (end users and companies) there
paying $20 (as an example), that's about $1B a year average as Larry
suggested. If you have 5.000.000 it's  only $100M. You at least have
some extra funds available that go DIRECTLY to developers. And as all
projects progress faster, you will have widespread adoption of OSS: I
will just be the natural step money-wise.

If 1B a year is not enough, then some projects will have to wait, or the
users base must grow (and it will grow surelly).  

The Linux Kernel and the core of OSS will have a fixed percentaje pumped
to them, so the ones that know what must be done are not bothered. Say
10% (or a different %). That's between $100M and $10M extra a every
year, and you can always count on it. If you don't need that much, then
you could resign the right. And if it's not enough, then it's much
better than nothing / "charity".

It works for Microsoft and they make crappy software which is also
propietary, closed, and expensive. Why wouldn't it work for OSS? You'll
be making everyone else in the world a favor. They collect most of their
revenues due to the Windows plataform. The kernel may provide that base
for all the GNU deck.

>From the end user perspective it would also make a lot of sense. Whoever
can't pay $20 year, doesn't know how to code, is not a student or a
charity organization may well thank everybody else's for they will be
able to use last year developement for free as in freedom and in beer.
If anyone is concerned about the impact $20 will have, as long as you
allow piracy (a la Microsoft), there's no problem. It may even help OSS.

Now the real concerns would be:

- Will the money be used wisely?
- Will money change the spirit of OSS and bring bad code with it?

I couldn't answer those. If the answer is negative, then OSS is better
off underfunded (as it is today). All for the love.

But what if it works out nice? You are indeed underestimating the
willing of non-coders to contribute. They will probably even be proud
members of the fGPL foundation.

Federico

On Sat, 2002-07-27 at 03:29, Alexander Viro wrote:
> 
> 
> On 27 Jul 2002, Federico Ferreres wrote:
> 
> > The fGPL scheme means you pay $10, and you get code in exchange. The
> > developer receiving the funds either buys food, a computer, a book, or
> > maybe even uses it for a nice trip with his/her family after their hard
> > work is done. Let me know when you find the pyramid!
> 
> Sigh...  That stops working at the same point where pyramid does - when
> number of recepients becomes a sufficiently large fraction of all potential
> participants.  It is not sustainable.
> 
> Simple math: Debian got several thousand developers.  Most of the packages
> contain both contributions from said developers and stuff from upstream.
> By very conservative estimates it's tens of thousands.  If your $10 is
> per package update - it's impossible to pay.  _Really_ impossible - on a
> reasonable system it will easily amount to $50 _daily_.  If it's $10 per
> year - count the number of installations and look how much it will give
> for one developer.
> 
> It won't work for the same reason why pyramids are unsustainable - when
> you have too many recepients, both go to hell.  And there _is_ too many
> for that to work.
>