Re: proposed improvements to PostgreSQL license - Mailing list pgsql-hackers

"We" (the Postgres steering committee) have discussed these issues for
months and months. We aren't trying to change anything, just reinforce
what we believe to be already the case. However, the path to do this
isn't perfectly clear to anyone; this is the first concrete proposal we
have had which does try to address the issues we believe are already
here whether we want them or not.

I'll bring them up farther down (and will probably forget and leave some
relevant pieces out).

> I think this is a bad idea for the following reasons:
> 1) It is trying to be a GPL in what it is trying to achieve without
> actually being well thought out. Any person who "submits" modifications
> must do so under the same licence. Submits to what or whom?

It is *not* trying to be GPL. It is trying to be BSD, while extending
liability protection to the current cast of developers, who are (I'm
pretty sure) not covered in any of the wording of the UCB-generated
license.

> 2) If the core team want to make sure modifications to the software are
> under the same licence then they should merely insist that any patches
> are accompanied by that same licence (i.e. the current licence). End of
> story end of problem. If you want to go any further than that you may as
> well go GPL.

The current license asks users to absolve the University of California
of any liability involving use of the Postgres source code. It does not
(currently) explicitly ask the same on behalf of the current developers
(including yourself ;)

> 3) You talk about how wonderful the BSD licence is, then you really
> change the whole meaning of that licence.

How?

> 4) What is this stuff about "tightening up of what the existing licence
> is supposed to do"? What do you think it is supposed to do? I think it
> is basicly an annoying artifact of UCB's legal team that happens to make
> the software virtually public domain. We might just as well get rid of
> all licences except that we're not allowed.

I disagree, though we don't know UC's motivations for sure. imho the BSD
license is intended to protect UC from "deep pockets" lawsuits, while
preserving some credit for the original design team and the institution
which made it possible.

The new wording is intended to continue to do exactly that, extending
the umbrella to cover developers with no connection to UC.

> 5) This "protection" for developers is a straw-man. I don't see, say the
> free-bsd developers worried about this. If Great bridge wants to
> distribute with extra disclaimers then go ahead.

It is being proposed as an addition to the Postgres development effort.
I'm sure that GB knows they could add anything they want to their own
product.

> 6) This is a very US-centric view of the world. Most of the developers
> are not in the US if the postgresql.org home page is correct. We don't
> care about the stinkin UCITA, we are not bound by and don't care about
> anything the State of Virginia may or may not say.

Good point. But the USA is the demon spawning ground for lawyers, and is
at the leading edge of aggressive new legal territory. That may change
eventually, but since 90+% of our federal legislative representatives
are lawyers (stats from memory, but it is a *high* number), that may not
change very quickly :(

> 7) I hope you're not thinking of bloating each and every source file
> with all that legalese.

No, afaik that is not considered necessary.

> 8) "To be integrated with the software in such a way that this license
> must be seen before downloading can occur".
> Umm, can all the laywers please just butt out? Every other open-source
> package in the universe just relies on a licence file in the home
> directory. You going to try and stop people downloading with clicking a
> licence agreement? How you going to handle mirrors? Or are  you not
> going to mirror any more? What about Red Hat el al?

Good point. Not exactly sure why this was suggested, but the American
courts are *full* of cases where the plaintif said that they "didn't
really know" something that should have been obvious.

> Point (8) makes me thing that this whole thing is the recommendation of
> some lawyer who is totally out of touch with the free software community
> but feels compelled to add a whole lot of disclaimers and so-forth
> because that's his job. Bottom line is it's not broke so leave it alone.

afaik "it's not broken" is true, for the free software community. And
part of my pleasure in contributing to Postgres is exactly because of
that general distain for legaleze and idiot-speak commercial agreements.

Postgres is starting to become a visible thing, and is going to be used
by people who don't know much about the free software movement. And
*I'm* within reach of the American court system, and *you* can
contribute code which could make me a target for a lawsuit. I'd rather
short-circuit that before the lawsuit, rather than asking for a donation
for my defense ;)

So the intent was, as stated, to *reinforce* what we already believe to
be true (including yourself). The recently-enacted UCITA law was
(afaict) intended to protect, perhaps wrongly imho, commercial software
companies from liability claims (I know that Oracle *claims* a whole lot
more for 8i than we do for Postgres, so why shouldn't they be held
accountable for what they claim?). But UCITA is a sharp tool which we
can use to protect volunteer software developers such as myself, and
you.

I (and *all* of the steering committee) had pretty much the same
reaction as you did at first. But some of us are closer to the US legal
system, and see what silliness it can generate, so came around to
thinking that there was something to be gained by license additions.

Regards.

                       - Thomas