A history of the Boost Foundation

From: Zach Laine (whatwasthataddress_at_[hidden])
Date: 2024-09-11 19:23:15

Dave Abrahams and Beman Dawes put up the initial money for BoostCon
(which has since been renamed to C++Now) in 2007 (or maybe 2006?).
This was something of a risk, since it was not certain how many people
would show up.  It turns out that the conference covered those costs,
but they were still personally responsible for the conference if they
wanted to put it on again the next year, or the year after.  Beman had
the idea that there should be an independent legal entity for this,
rather than two individual people.

I forget the exact chronology, but pretty early on in the conference
history, we ended up with The Boost Steering Committee, under the
Software Freedom Conservancy (SFC).  That is, we called it "The Boost
Steering Committee", but there was no such legal entity.  The legal
entity that represented the BoostCon conference and Boost in general
was SFC.  In terms of legal entities, "The Boost Project", or "The
Boost Steering Committee" was just a single division within SFC.  This
is something they reminded us of from time to time.

At the time, Boost had no real expenses.  The "wowbagger" server that
we use to host the (now, legacy) website, and on which the releases
are built, was hosted at Indiana University.  The downloads of the
build Boost releases were hosted at Sourceforge (I think; anyway,
wherever it was, it was free).  They later moved to other places,
ending up finally at JFrog.  They're no longer hosted for free -- see
below.

BoostCon made more money than it spent for a lot of years.  It
accumulated some reserves, and people would sometimes ruminate on what
we might spend all that sweet, sweet cash on.  No large expenditures
were ever taken on though.

The SFC started very small.  We were project #3 I think (maybe #2?
something like that).  Over a few years, it took on more and more
projects, until Boost was one of 20 or 25 SFC projects (again, I don't
remember the exact numbers). SFC had a small staff then, and probably
still does.  As they grew, it became increasingly difficult to get
anything done.  They were overworked and understaffed, so this is not
some failure on the part of the SFC so much as a logistical
inevitability.  One problem we had was getting them to pay vendors for
services used by the conference.  IOW, we would use some service, the
vendor would send us a bill, and then we would send SFC like a
thousand emails trying to get them to pay the bill.  It was a
nightmare to deal with.  Jon Kalb, who was the conference chair at the
time, asked the Steering Committee if we'd be open to breaking free of
SFC, and forming our own non-profit organization -- a separate legal
entity.  Everyone agreed, Jon did a bunch of research and paperwork
(thanks again, Jon!) and we formed a new 501(c)(3) US corporation
called The Boost Foundation.  This is why the
Steering-Committee-under-SFC became the Boost Foundation Board.  This
is also why the name changed -- we needed a new name for the new legal
entity.  There was no secret agenda; there was no smoke-filled room.
There was, however, pragmatism.

Sometime around then -- once more, I don't remember the exact timeline
-- we started losing all our free stuff.  Someone at Indiana
University realized that randos (that's us, Boost) were running a
whole server in their lab, and asked us to remove it.  So then we had
to pay for hosting wowbagger.  Some time later, the download hosts
stopped giving away bandwidth for free as well.  No problem, right?
The conference makes money, Boost now needs money, so let's use the
conference money for the Boost expenses that now have to be paid.
(Note that this is something the SC/Board took on as it came up; it
was not part of the original mission of the SC/Board, because there
were no significant Boost infrastructure expenses previously.)  The
problem is that the conference makes less and less every year, due to
a combination of changes in the way the venue charges (it all went up
a lot), inflation, the expenses from the covid-cancelled year, etc.

During all this, we kept things running as best we could.  Eventually,
the downloads became too expensive, and the C++ Alliance offered to
pay for those.  So now they do.  The Foundation only pays for hosting
wowbagger right now.

Hopefully that sets the stage.  To me, the important take-away is that
the Foundation has tried to "keep the lights on" as I like to say,
without bothering anyone on the list about it.  The Board does more
than just Boost infrastructure stuff, such as organizing the C++Now
conference, and more recently, paying the hosting fees for the Beman
Discourse server.

Sorry this is so long. :)  I may have some of these facts slightly
wrong.  Like, did BoostCon actually start in 2006?  I couldn't figure
out definitively which year it was, even though I was there.  But I'm
much more certain about the major events, and the reasons for things.
I also left out most of the non-asset-stewardship aspects of the
Steering Committee/Foundation, since they are mostly not germane to
the current review.  I nevertheless refer to one such aspect below.

In light of the history above, here are some points I'd like to make
regarding things I've seen said on this list about the Foundation:

1) The Board does not impose its will on the developers on this list.
It does sometimes make recommendations or official "calls" of some
kind, like saying we should use CMake.  It doesn't actually have the
ability, as the Boost Foundation Board, to enact such changes.  This
is not a matter of restraint.  It simply can't make concrete changes;
only the Boost developers can.  Also, there never has been, and
doesn't appear to me that there ever will be, any appetite for
meddling in Boost development, or forcing developers to do things.

2) The Board is self-selected, in the sense that the Board nominates
and votes on Board seats.  Multiple Boost authors are on the Board:
Peter Dimov, Glen Fernades, Jeff Garland, and me.  There are others
who are only there for the conference business.  There are others, who
joined more recently, interested only in Beman.

3) I've read on this list about the desire to "get back" to the Boost
Steering Committee way of doing things.  This does not make sense.
There is no difference between the Boost Steering Committee and the
Boost Foundation, except for the name.  Sure, there is a change in
membership over time.  That would have happened with or without the
name change.  Perhaps this is a desire not to have the current Board
membership?  If so, please get involved.  Asking a volunteer project
to do things differently, without volunteering to help yourself,
doesn't usually work out.  Which brings me to:

4) The Board is a volunteer project, like Boost is.  As such, people
work mostly on what they care about, and don't really do much for
things they don't care about.  So when I read, "The Board is not
communicating enough," I think there are a couple of things to
consider.  4a) Do we expect volunteers to do the stuff they want to
work on, and then communicate that they did that to someone else that
was not working on that same thing?  Do we do that with our Boost
libraries?  I might communicate changes, including bug fixes, in
release notes.  I do this because it affects others.  But if I make a
non-functional change, who do I tell, and why?  Note that, when there
has been a loss of service, someone has mentioned it on the list right
away, as far as I know.  Do we also want people to mention when the
Board took some action that kept a loss of service from happening?
Moreover, I'm not sure what transparency is lacking, given that: 4b)
The board publishes minutes of every meeting, and you can come to the
meetings if you're interested.  Except for occasionally needing to
handle a sensitive topic that we do not minute due to its sensitivity,
these are not closed meetings.  For example, our latest Board member
was asked to join the Board because he kept showing up to these open
meetings out of his own interest, and would take on tasks from time to
time. As you all know, this is how pretty much all open
source/volunteer projects work.  In short, I think if a Boost user
wants to know the details of how Boost works, they subscribe to the
Boost mailing list.  If a Boost developer wants to know the details of
how the Boost Foundation works, they read the minutes and/or come to
the meetings.

Zach