On Thu, 16 Dec 2004 13:37:12 -0800, Michael Latta <lattam at mac.com> wrote:
> This brings up something I was thinking about recently.  Is it kosher
> to charge for enhancements to Squeak?  Even if you adhere to the letter
> of the license, is it socially acceptable to this community.  What if
> the next release of MC cost $5, what if a JIT cost $25.

I suspect the result of charging $5 for the next release of MC would
be that almost everyone would simply continue to use the current
release, or add whatever new features they wanted themselves.  This
isn't even because people care about the $5, but because it's less
hassle and more peace of mind to use something that's free (especially
if there was any attempt to circumvent pirating; nobody wants to enter
a serial number every time they load MC into an image).  I could be
wrong, but my guess is that for a tool to gain widespread adoption in
the Squeak community, the code itself needs to be available freely;
anything else will get routed around.

Something like a JIT is a slightly different story, because it doesn't
really need widespread adoption to be effective.  There's no network
effect with JITs; the fact that nobody else in Squeak is using it
doesn't make it significantly less useful to someone who is.  So in
theory you could license a JIT for a large amount of money to a very
small number of customers who really cared about their Squeak
application going faster.  This is how Zend made money (or tried to)
from PHP.  The problem is that in Squeak's case, this number may be <
1.

And, come to think of it, there is a potentially interesting network
effect from having a free JIT for Squeak, which is that you may
attract more people to Squeak itself - if you could say "Squeak is as
fast as Java" and mean it, Squeak might grab some market share it
wouldn't have otherwise, which would, among other things, increase the
profile/reputation/employability of the JIT implementor.  This might
be worth more in the long run than the (<1 ?) licenses you would have
sold.

So I'm somewhat skeptical about selling licenses for Squeak tools
(this is very different, of course, from selling licenses to endusers
for things built on top of Squeak, which is very worthwhile).  But it
does seem like there should be some way of getting funding to the
people that are doing interesting work in the community.  A few
options:

- bounties.  This has already been mentioned.  For example, people
might pledge to pay $500 for a VM that increased the send speed by 5x
or more.  If enough people made that pledge, a JIT implementor might
become more motivated.  This probably works best for things like
performance that can be (somewhat) objectively measured; I'm not sure
what kind of a bounty someone might offer on version control features,
though I'd be happy to find out :).
- prioritized support.  As you'd expect, people that retain me for
consulting, or negotiate a support contract with me, get their emails
and bug reports (say, for MC) addressed somewhat quicker and more
fully than anyone else.  It might be possible to formalize this
somewhat and have, for example, a "registered user" mailing list with
a yearly subscription fee, where messages were guaranteed to be
treated as high priority by the tool implementors.
- donations.  Either to individual projects (a paypal link on each
SqueakMap entry?), or to a central foundation; with this last,
probably the most effective thing would be for the foundation to then
fund individual Squeakers for a few months or a year at a time,
letting them focus full time on contributing to Squeak.  This is,
IIRC, how development on the Perl 6 compiler happened for example.

Some of these things have been tried before - there's a page somewhere
on the swiki with a list of bounties, for example.  But since nobody
has really come up with a successful model yet, it's worth continuing
to talk about.

Avi