Summary time. I have had several private responses, as well as some copied
to this list. A number of helpful bits of advice have come forward. But
first just to correct some frequent mistakes:
1. Most people who felt a need to write about the GPL aspects
misunderstood how the GPL works. The FSF has a good summary at
http://www.fsf.org/philosophy/selling.html, which specifically
encourages people to make as much money as possible selling GPL
source and/or binaries. There is no need to distribute source with
binaries, but the source must always be available.
In this case, the best way I can think of making the source
available is to do the port well enough that it is accepted into
the base Exim distribution. There can be no ambiguity then.
It is also possible that I wasn't very clear about my understanding
of the GPL the first time I wrote, but hopefully this is better.
Nevertheless, it is surprising how many people simply don't follow
how the GPL is designed. Selling GPL'd binaries is a very good
thing for free software, and it isn't terribly easy to make money
doing it for Unix-like platforms - ask any Linux or FreeBSD
CD seller how much money they make from selling, say, Samba :-)
2. I would hope that neither this nor any other Exim packaging project
ever misrepresented the Exim name. It would not be the
intention of this notional project to get a free ride from the Exim
name. Several people suggested that Philip protect the name and I think it
is a very good idea, if he hasn't already. The approach
taken by many projects is that anything that is part of the main
source tree can take the name, and anything else must differentiate
itself clearly by using a different name.
3. Some seem to think I somehow like Win32, or think it makes a good
platform for an MTA! No, this isn't close to true :-) However
there are plenty of people for whom it is true, and who can
potentially be used to fund work on a better Unix MTA if
they are willing to fork() out cash for a Win32 version.
A consolidated summary of comments I got back, with my responses:
a) "This was an RFC, not an FAQ."
Ok, RFC then. Whatever :-)
b) "Wrong group of people targetted."
I don't think so, because the replies (especially the private
ones) have told me a lot about what needs to happen to make sure
people understand why this is not only legal, but also a good thing
for Exim. If some port of Exim is doing things that exim-users think
stinks, that won't be good for either that port or for Exim. I have
had a lot of helpful feedback that would guide how this notional
project would proceed.
c) "There will be no market"
I am quite sure there won't be much of a market among existing Exim
users :-) What other market there is depends on many other factors.
d) "There will be a market"
Other people seemed to feel like there isn't much quality choice in
MTAs for those who must or wish to use Windows, and that a working
Exim for Win32 might be a lifeline for Windows users desperately
looking for solutions.
e) "The value-add bits are fair enough and laudable"
Well at least many people didn't seem to mind about the frilly
add-ons, even if some of them were worried about potential GPL
violations. This much seems to go down well.
Nobody mentioned the potential to increase the size of the Exim-using
community, nor of the possibilities of easy conversions to Unix-based
Exim. Maybe nobody cares.
Nobody volunteered any work done on this. Philip knows of two Win32
projects that have sunk without trace, so it is probably fair to say that
there is no work to build on.
--
Dan Shearer
Open Source Manager
Mob: +61 411 49 1800
Tel: +61 8 8130 3104
dan@???
