> As just said, this is a good thing from the point of view of > encouraging participation and commercial success.
Aslong as the open > source version of postgresql remains a well-designed, solid product it > behooves any commercial
distributorto aid in its maintenance rather > than take on the whole thing. Ideally, they will contribute any fixes
>they make so that all can benefit and perhaps more importantly so they > don't have to maintain the separate fixes
anylonger.
this assumes an ideal world.
Not entirely. It really assumes that people do a cost-benefit
analysis and recognize that the cost of maintaining a separate
distribution with patches, etc. of something as complex as PostgreSQL
generally far outweighs the benefit of having a global network of
programmers do it for you. Simple economics, not idealism.
look at BSD itself and how it has fragmented, or the plethora of unices and how the forking there nearly closed the
dooron wide spread use. the fact of the matter is, eventually someone/someidiot feels slighted by the community or
feels their ideas are better, no matter WHAT. so they go their own way. they decide to take the BSD'd source and
wanderoff to their little corner of the world, taking some of the developers with them.
Look also at Linux and how it has fragmented despite the GPL. There
is not one Linux release, at all, despite the common usage of the term
"Linux" as if that referred to a single entitity. In contrast, the
term "NetBSD", for example, refers to exactly one public release, a
standard and well defined entity that can be readily duplicated.
Indeed, there are likely many more versions of Linux in standard
distribution and use (at least 17 in a recent count) than ALL of the
publicly-released *BSD OSes combined (3).
whilst BSD is less restrictive, its so unrestrictive that it allows people to set up barriers to the furtherance of
thesource.
But as long as the publicly available source remains a viable
enterprise (e.g., people see strengths in it and are willing to
contribute time rather than pay for a commercial version) it doesn't
matter if there exist other versions that have been commercialized.
The issue is whether or not PostgreSQL should ALLOW release of binary
versions or not. There are many viable situations in which it is not
feasible or desireable to ship source code, even if the product is
identical to the public source. For example, clients may not care and
may not want to be burdened with it. Or the database may be embedded
within something else in a streamlined system for which there is no
space for source. With the GPL, the producer is required to either
ship the source or maintain all the relevant copies for at least 3
years. That is quite a burden when no one cares to have the source.
In this case the GPL ultimately restricts what can be done in an
economically realistic manner and can lead to stagnation.
In the interests of maximizing the potential marketplace for
PostgreSQL, it seems like the BSD license is in fact superior.
finally, does the BSD liscence further the possibility of commercial adoption? well. look at Apache. apache has
commercialproducts built on top of it. PHP4 w/commercial optimizer leaps to mind. and commercial acceptance (e.g.
bundlingwith other packages that are closed, open, free, for sale, etc) is quite high.
Both Apache and PHP (both 3 and 4) have BSD-style (not GPL) licences
(though you can choose to abide by the GPL for PHP4, if you want, you
are not required to). Hence, commercial vendors can release binary
products built upon either of these without also releasing the source.
That is likely the main reason these products have been adopted in the
commercial world (given the prerequisite that they are both extremely
high quality products to start with).
the ability to sell a core product directly does NOT create commercial success. instead, i feel it encourages
forkingand the removal of that product from the user base. which is us. remember: greed and pride cause people to do
stupidthings.
Clearly the ability to sell something does not create a success. Nor
does it necessarily encourage the forking you seem to think is
inevitable. Your examples above both counter this claim.
that being said.... is it possible to GPL postgresql? probably not, i'd imagine. how much of the original Berkley
sourcecode is there left?
Nor is it desireable, I would argue.
Cheers,
Brook
