Re: [HACKERS] proposed improvements to PostgreSQL license

>One issue that has always been a source of uncertainty - I think for
>all of us - has been the license under which PostgreSQL is
>distributed.

Only an issue of uncertainty when people talk about changing it.

As a company who wants PostgreSQL to remain in the public domain, I would
prefer to see it go GPL; this effectively prevents another company coming
along and swallowing the major developers as a means of stifling further
development. This latter tactic would probably never work, but it would be
extermely disruptive.

As a developer who likes to be able to make a buck from their work, I like
the idea of licencing my code for free use within, and only within, the
free, open-source version of PostgreSQL.

It seems to me that if you want to do all that you say, this is probably
the way to go.

>We've also found, through some rather extensive market
>research

Out of curiosity, who with & where?

>I don't want to re-start that debate here

Sorry.

> - the
>consensus in the PostgreSQL community over the past few years seems
>to be that the Berkeley style license is best suited for the
>continued development of PostgreSQL.

News to me. Perhaps an informed survey of the current PostgreSQL user base
might be in order (perhaps each party - BSD, GPL, and other - can put their
case in 500 words or less?). I would be very interested to know the
outcome. It seems to me that the camps are highly divided, perhaps the
majority prefer BSD, but I suspect that they all agree that the status-quo
is acceptable. Once you seek to change the agreement to another
non-standard agreement, I think you will be opening a can of worms. You
will at least have to persuade all past contribitors that your new version
is better than their favorite model and BSD.

>protecting the developers who worked on the code, and ensuring that
>the code stays open source in perpetuity.

No; this is what companies seeking to develop private versions of the
software want. As it stands, all developers who have contributed to PG over
the years still own the copyright of their code, and can withold the
license to use it. Unless they have assigned that copyright somewhere else,
or signed the sort of agreement you propose. Releasing code to the PG
community provides an implied license to the PG community. It does not
necessarily grant rights to others, unless specifically stated in the code.
This is why you have:

"Any person who contributes or submits any modification or other
change to the PostgreSQL software or documentation grants
irrevocable, non-exclusive, worldwide permission, without charge, to
use, copy, further modify and distribute the same under the terms of
this license"

which just dimishes the rights of the developer make a buck if someone else
uses their code as part of a commercial version of PG.

I have just submitted a chunk of code to the project, and the notice in the
code says, basically, it is free to be used in any way. But I would have
*much* preferred it to have been under the GPL; then when someone fixes my
code, or improves it, I know that I will get to see (and use) the changes.

>Unlike other open source licenses (GNU, Mozilla,
>Interbase), the original Berkeley license does not take into account
>that over time a lot of different individual developers and perhaps
>some corporate contributors, would have individual copyrights on
>substantial portions of the code.

It does not need to in the sense that while it remains public and in it's
original form, an implied license is granted. Otherwise the notice that the
developer placed in the code holds, and failing that, whatever the
governing law for the country/state of origin applies. Generally this means
that the developer owns the copyright, but gives others rights to use it in
the public version of PG.

>This deficiency has two adverse affects. First, the contributing
>developers are not afforded the protection of the exculpatory
>language in "bold face."

I agree this is a problem. Developers should be warned to always add some
kind of text like this to all their public code - not just PG.

>Second, and admittedly of
>significant importance to Great Bridge, the commercial proliferation
>of PostgreSQL could be hindered if business users are concerned that
>the license might not cover the substantial additions and
>improvements made to the code over the last few years.

They will be covered so long as they release their developments back into
the community, I think.

>In developing the new language, the resources of two intellectual
>property law firms, one East Coast and one West Coast, were tapped.

What about European (east and west), Japanese, and Australian?

>No discussion of this type should be without controversy, so I throw
>you the following red meat: the choice of state law has been
>selected to cause the application of the Uniform Computer
>Information Transactions Act (UCITA) to the usage of the software.
>
>(Pause for outrage to subside.)

What is it? Does it even apply to me? [I am awaiting advice on this from my
IP lawyer]

>I submit that your efforts should be exempted from any potential
>liability.

This is good. We can add the extra <bold face> paragraph.

>Two states have adopted UCITA - Virginia and Maryland.

As a matter of interest - do Virginia & Maryland have a reputation for
being forward looking in their law making, and being protective of the
right of individuals over companies and government? I ask this because I
know very little about individual US states...

>had to be contributed back to the public use, would be poorly
>received and would possibly discourage the use by businesses who
>might want to make certain improvements for their internal use

How about allowing developers the choice, as they have now?

----------------------------------------------------------------
Philip Warner | __---_____
Albatross Consulting Pty. Ltd. |----/ - \
(A.C.N. 008 659 498) | /(@) ______---_
Tel: (+61) 0500 83 82 81 | _________ \
Fax: (+61) 0500 83 82 82 | ___________ |
Http://www.rhyme.com.au | / \|
| --________--
PGP key available upon request, | /
and from pgp5.ai.mit.edu:11371 |/