Subject: Re: Open Source funding?
From: Ian Lance Taylor <ian@airs.com>
Date: 18 Feb 2005 14:28:02 -0500

Lajos <lajos@galatea.com> writes:

> 5) the volunteer non-profit model built around a product (Apache,
> ObjectWeb) with unpaid or paid volunteers
> 
> 6) the volunteer non-profit model built around a product (Apache) with
> only unpaid volunteers (SourceForge-type projects)
> 
> ...
> 
> What interests me most are the #5 & #6 models. Neither of these are VC
> funded, because they have no profit motives. However, these are often
> the models in which some of the best open source software is
> developed. Hence I'm interested in exploring the question of the
> long-term viability of these projects. So far, Apache does very well
> because copmanies using open source software are willing to have
> employees spend paid time working on that software. How long will this
> last? What are the benefits realized by these companies who contribute
> employee time? Will these benefits continue long term? Perhaps Brian
> can help me out here ...
> 
> With the #6 projects, however, I frankly wonder how long they can
> last. The accounting project I referenced in my original post is in
> this group. A cyncial predication would be that this project will
> eventually be muscled out by a VC-backed competitor along the #1/#2
> model. This I don't like. I want to see projects of the #5/#6 models
> survive. No, they are not the vaunted "professional open source" but
> they are responsible for some darn good software. In my example
> project, the software and founders have really done some excellent
> work - they just don't have the muscle to bring the project to its
> completion.

You seem to be applying market considerations to projects which are
not market driven.  It's like saying that volunteer-run soup kitchens
are going to get muscled out by professional restaurateurs.  A few
unpaid free software projects are like little companies, but most are
more like a club or a pickup sports league.  They can't be replaced by
companies because they aren't playing the same game.

That said, free software projects are of course competing in the
market of ideas.  In that market, if somebody produces a better free
accounting package, that is good for (almost) everybody, no matter
where it comes from.  Whether it is bad for the accounting package
which people stop using depends on the motivations of the people
writing it.  If their goal is to get it widely adopted, then it is bad
for them.  But it seems clear from looking at sourceforge that many
people work on free software projects for other reasons.

Further, if the original accounting package is good, then it is likely
that a company which wanted to produce a free accounting package would
improve the existing one, rather than writing their own from scratch.
They might work with the original authors, or they might find it more
expeditious to work separately.  To use your words, they provide the
muscle which the original authors may lack.  Again, almost everybody
benefits.  (Although I do want the question the notion of lacking
muscle.  What are we talking about here?  More programmers?  A bigger
Q/A team?  GUI designers?)


On a different topic, you ask above whether companies will continue to
let their employees work on free software projects.  The answer is
that they will as long as the derive some benefit from it.  The most
obvious benefit comes when they use the package themselves.  Many
companies use Apache, but few companies sell Apache.  A few of those
using Apache no doubt contribute patches of their own written by their
paid employees.  Those patches are to the benefit of the contributing
company, as they get a better product without the difficulty of
maintaining and updating their own separate fork.


> Frankly, I don't trust the industry (yes, I can hear the free
> marketers on this list gasp). If we just rely on what gets funded, or
> what names the Wall Street companies see in the open source
> marketplace, a ton of really good stuff is going to go down the
> toilet. And with that goes a lot more that is part of open source as a
> movement, I think.

As I implied above, I don't see how that can happen.  Good free
software will not be lost.  Bad free software will be.


> So, back to my question: does anyone have experience with projects of
> the #5 or #6 model getting funded by grants or any kind of serious
> donations?

There are many projects, now funded, which got started with no
funding.  The Linux kernel is an obvious example.  So is the GNU GCC
compiler.  Historically speaking (if I can use the term "history" to
cover a time span as short as the free software movement), the notion
of a free software project having been funded from the start is a
relatively new one.  Aside from academically funded projects like BSD
Unix, probably the first such project was Mozilla; recall that freeing
it up was a desperation move, and that for a few years it was the only
such project.


> What would people suggest to a really good #6 model-type
> project that doesn't have paid employees working on it but which has
> brilliant work behind it? Should they try to turn themselves into a #1
> or #2 type model and seek VC funding?

The issue of VC funding in particular is a complicated risk/reward
calculation.  The money comes with big strings attached, and anybody
seeking VC funding should understand those strings in detail.  (I was
a co-founder of a VC backed company, now defunct).

Of course, all money comes with strings attached (except outright
grants, of which a few do exist now in the free software world).  So
any programmer considering seeking funding of some sort should be
asking "what will it do for me" and "what will I have to change in
order to get it."  There is no one answer.  It really depends on the
people involved, and what their personal goals are.

If the goal is the widest possible adoption of the free software
package, then VC funding is only appropriate if there is a clear path
to deriving revenue from some secondary aspect.  An example here would
be Macromedia Flash, in which the player is free, and could easily be
free software, but the program which creates Flash applications is
not.  Otherwise, I think the best way to get wide adoption of a free
software package is networking, the Slashdot effect, making the
software clearly better than its competitors, etc.  Not so different
from a company trying to get wide adoption of a program, actually,
except that the free software developers have the benefit of low cost.

Ian