FreeBSD funding [was: Re: Benchmark (Phoronix): FreeBSD 9.0-RC2 vs. Oracle Linux 6.1] Server

I have slightly reordered your email in my reply, in order to put the
most important item last.

On Fri, Dec 23, 2011 at 12:01:33PM +0100, O. Hartmann wrote:
> I'm still with the system, although I desperately need scientific grade
> compilers or GPGPU support.

Your use-case, while valid, is clearly not the use-case that most of
the committers working on FreeBSD face.  But see below.

> And, I dare to put some critics herein! Since I see that FreeBSD is
> "free", why not trying to make it better and more towards perfect?

Everyone wants the product to improve.  The question is, what is
achievable with the current committers?  That's where you see the
pushback and frustration from the current committers.

> Look at FreeBSD and the problem of how well sysctls and their
> working are documented. It needs to be fixed.

There's no argument that some of the FreeBSD documentation is stale.
We do, however, have one committer (eadler_at_) who has been trying to
move the sysctl documentation forwards.

Participation from the wider community is key.  Although sending PRs
does not guarantee things will get fixed, it's currently the best way
that we have.

> Well, as far as I know, the FreeBSD project is funding people doing a
> certain work! So, the implied opposite, FreeBSD is developed "free"
> isn't true.

So here's the key point of your email IMHO, and the key misunderstanding.

First, let me nit-pick the legalities.  The FreeBSD Foundation
(http://www.freebsdfoundation.org/) is a US non-profit that does fund
some activities, and that's what I'll talk about here.  "The FreeBSD
Project" is the collective term for "all of the committers and developers"
and is not an "entity" for US legal purposes.

Second, the disclaimers: I am not a member of the FreeBSD Foundation
Board of Directors, so I am not speaking for them.  I have also directly
benefited from Foundation funding (both travel, and via equipment they
bought for portmgr), so am hardly unbiased.

Now on to the gist of that matter.

As a US non-profit, the Foundation is required to post its financial
information to the public, and it does on its website:

  http://www.freebsdfoundation.org/documents/Budget2011.pdf

You'll see here that the total budget for 2011 is $400k (USD).

This, frankly, is miniscule.

The largest line item for 2011 is $125k for project funding, which has
gone towards 9 different projects (see
http://www.freebsdfoundation.org/activities.shtml).  For comparison,
keep in mind that a commercial developers' salary in the US is upwards
of $100k/yr.

Even with this being a substantial increase from 2010's $83k, these
numbers are tiny comared to "real-world" budgets.

The projects that were sponsored were primarily networking-related, but
also the GEM/KMS/DRI project, jails, the libc++ replacement, and clocks.
I've listed those in the order that I think the most consumers of FreeBSD
will be affected by.  Note the absence of any work towards performance,
schedulers, compilers, or numerical analysis.  With a $125k budget, you're
simply not going to see those on the list.

The other notable line items are: hardware purchases (explicit disclaimer:
portmgr has been one of the primary beneficiaries); conference sponsorship;
conference travel; and salary for one employee to try to help coordinate all
the above.  Legal fees (things involving trademarking and licensing issues)
takes up most of the remaining.

I can't figure out the Linux Foundation's budget from their website, but I
can tell immediately that their budget is a great many times more than $400k.

Summary: on a fraction of the budget that Linux has available, we _nearly
keep up_.  I can't imagine what we could do with comparable funding.

So, for everyone who thinks we are being "well funded", here's your reality
check.

And please note that the Foundation is in its year-end fund drive, too.

Thanks for listening.

mcl