Thread: Patent warning about the Greenplum source code

Some of you might have seen that the Greenplum database source code has
been published:
https://adtmag.com/articles/2015/10/28/greenplum-open-sourced.aspx

under the Apache 2.0 license:
http://www.apache.org/licenses/LICENSE-2.0

The source code has known patents owned by Pivotal/Greenplum.  The
license has a patent grant clause:
3. Grant of Patent License. Subject to the terms and conditions of thisLicense, each Contributor hereby grants to You a
perpetual,worldwide,non-exclusive, no-charge, royalty-free, irrevocable (except as stated inthis section) patent
licenseto make, have made, use, offer to sell,sell, import, and otherwise transfer the Work, where such licenseapplies
onlyto those patent claims licensable by such Contributor thatare necessarily infringed by their Contribution(s) alone
orbycombination of their Contribution(s) with the Work to which suchContribution(s) was submitted. If You institute
patentlitigationagainst any entity (including a cross-claim or counterclaim in alawsuit) alleging that the Work or a
Contributionincorporated withinthe Work constitutes direct or contributory patent infringement, thenany patent licenses
grantedto You under this License for that Workshall terminate as of the date such litigation is filed.
 

However, while the license defines and uses "Derivative Works", it does
not mention that in the patent grant clause.  I assume this means that
patent grants do not apply to derived works, meaning if code or ideas
were moved from Greenplum to Postgres (which is not Apache 2.0
licensed), it would not have a patent grant. I talked to Greenplum staff
about this a few months ago and they did not dispute my analysis.

Therefore, I caution people from viewing the Greenplum source code as
you might see patented ideas that could be later implemented in
Postgres, opening Postgres up to increased patent violation problems.  I
am also concerned about existing community members who work for
Pivotal/Greenplum and therefore are required to view the patented source
code.  The license issue might eventually be improved by
Pivotal/Greenplum, but, for now, I think caution is necessary.

Of course, never mention known-patented ideas in any community forum,
including this email list.

--  Bruce Momjian  <bruce@momjian.us>        http://momjian.us EnterpriseDB
http://enterprisedb.com

+ As you are, so once was I. As I am, so you will be. +
+ Roman grave inscription                             +

Hi,

I don't really want to discuss patent issues publically.

On 2015-10-30 04:47:35 -0400, Bruce Momjian wrote:
> However, while the license defines and uses "Derivative Works", it does
> not mention that in the patent grant clause.  I assume this means that
> patent grants do not apply to derived works, meaning if code or ideas
> were moved from Greenplum to Postgres (which is not Apache 2.0
> licensed), it would not have a patent grant. I talked to Greenplum staff
> about this a few months ago and they did not dispute my analysis.

The easiest thing would be to dual-licensce the code such contributions
to postgres. That sounds quite possible to me.

> Therefore, I caution people from viewing the Greenplum source code as
> you might see patented ideas that could be later implemented in
> Postgres, opening Postgres up to increased patent violation problems.  I
> am also concerned about existing community members who work for
> Pivotal/Greenplum and therefore are required to view the patented source
> code.

Issues around this are much larger than patents. Any contribution done
under employment has such risks. That's why the kernel has the
signed-off-policy.

Check the section about signed-off-by in
https://www.kernel.org/doc/Documentation/SubmittingPatches
and simpler
https://ltsi.linuxfoundation.org/developers/signed-process


Greetings,

Andres Freund

On Fri, Oct 30, 2015 at 09:56:48AM +0100, Andres Freund wrote:
> Hi,
> 
> I don't really want to discuss patent issues publically.

While we don't want to discuss patented ideas, the patent terms are an
imporant topic here.

> On 2015-10-30 04:47:35 -0400, Bruce Momjian wrote:
> > However, while the license defines and uses "Derivative Works", it does
> > not mention that in the patent grant clause.  I assume this means that
> > patent grants do not apply to derived works, meaning if code or ideas
> > were moved from Greenplum to Postgres (which is not Apache 2.0
> > licensed), it would not have a patent grant. I talked to Greenplum staff
> > about this a few months ago and they did not dispute my analysis.
> 
> The easiest thing would be to dual-licensce the code such contributions
> to postgres. That sounds quite possible to me.

Yes, but once they get contributions from outside, that is much harder to
add.

> > Therefore, I caution people from viewing the Greenplum source code as
> > you might see patented ideas that could be later implemented in
> > Postgres, opening Postgres up to increased patent violation problems.  I
> > am also concerned about existing community members who work for
> > Pivotal/Greenplum and therefore are required to view the patented source
> > code.
> 
> Issues around this are much larger than patents. Any contribution done
> under employment has such risks. That's why the kernel has the
> signed-off-policy.
> 
> Check the section about signed-off-by in
> https://www.kernel.org/doc/Documentation/SubmittingPatches
> and simpler
> https://ltsi.linuxfoundation.org/developers/signed-process

Yes, this does expose a missing part of our existing process.

--  Bruce Momjian  <bruce@momjian.us>        http://momjian.us EnterpriseDB
http://enterprisedb.com

+ As you are, so once was I. As I am, so you will be. +
+ Roman grave inscription                             +

On Fri, Oct 30, 2015 at 04:47:35AM -0400, Bruce Momjian wrote:
> Therefore, I caution people from viewing the Greenplum source code as
> you might see patented ideas that could be later implemented in
> Postgres, opening Postgres up to increased patent violation problems.  I
> am also concerned about existing community members who work for
> Pivotal/Greenplum and therefore are required to view the patented source
> code.  The license issue might eventually be improved by
> Pivotal/Greenplum, but, for now, I think caution is necessary.
> 
> Of course, never mention known-patented ideas in any community forum,
> including this email list.

I just found out that Citus Data has patent applications pending, so
viewing Citus Data source code has the same problems as Greenplum.

--  Bruce Momjian  <bruce@momjian.us>        http://momjian.us EnterpriseDB
http://enterprisedb.com

+ As you are, so once was I. As I am, so you will be. +
+ Roman grave inscription                             +

On Sun, Nov  1, 2015 at 01:27:13AM -0500, Bruce Momjian wrote:
> On Fri, Oct 30, 2015 at 04:47:35AM -0400, Bruce Momjian wrote:
> > Therefore, I caution people from viewing the Greenplum source code as
> > you might see patented ideas that could be later implemented in
> > Postgres, opening Postgres up to increased patent violation problems.  I
> > am also concerned about existing community members who work for
> > Pivotal/Greenplum and therefore are required to view the patented source
> > code.  The license issue might eventually be improved by
> > Pivotal/Greenplum, but, for now, I think caution is necessary.
> > 
> > Of course, never mention known-patented ideas in any community forum,
> > including this email list.
> 
> I just found out that Citus Data has patent applications pending, so
> viewing Citus Data source code has the same problems as Greenplum.

Actually, it might only be their closed source software that contains
patents, i.e. not pg_shard.  I will check and report back when I can
unless someone else reports here first.

--  Bruce Momjian  <bruce@momjian.us>        http://momjian.us EnterpriseDB
http://enterprisedb.com

+ As you are, so once was I. As I am, so you will be. +
+ Roman grave inscription                             +

On 10/31/2015 11:47 PM, Bruce Momjian wrote:
> On Sun, Nov  1, 2015 at 01:27:13AM -0500, Bruce Momjian wrote:
>> On Fri, Oct 30, 2015 at 04:47:35AM -0400, Bruce Momjian wrote:
>>> Therefore, I caution people from viewing the Greenplum source code as
>>> you might see patented ideas that could be later implemented in
>>> Postgres, opening Postgres up to increased patent violation problems.  I
>>> am also concerned about existing community members who work for
>>> Pivotal/Greenplum and therefore are required to view the patented source
>>> code.  The license issue might eventually be improved by
>>> Pivotal/Greenplum, but, for now, I think caution is necessary.

Do let me point out that *code* isn't patented.  *techniques* are.  So
those techniques are patented whether or not you read the code.  It's
just that if you read the code, copy the technique directly, and put it
in Postgres, that's considered "willful" instead of "innocent"
infringement and the penalties are different.  Its effect on our project
is the same, though: we have to rip out the code in a hurry.

Maybe we should just relicense PostgreSQL as Apache and cover all of the
patent issues ;-)

>>>
>>> Of course, never mention known-patented ideas in any community forum,
>>> including this email list.
>>
>> I just found out that Citus Data has patent applications pending, so
>> viewing Citus Data source code has the same problems as Greenplum.
> 
> Actually, it might only be their closed source software that contains
> patents, i.e. not pg_shard.  I will check and report back when I can
> unless someone else reports here first.

I will ask Citus Data for an official statement.  I will point out that
cstore_fdw is Apache-licensed, which also contains a patent grant.

-- 
Josh Berkus
PostgreSQL Experts Inc.
http://pgexperts.com

On Sun, Nov  1, 2015 at 12:12:48PM -0800, Josh Berkus wrote:
> Do let me point out that *code* isn't patented.  *techniques* are.  So
> those techniques are patented whether or not you read the code.  It's
> just that if you read the code, copy the technique directly, and put it
> in Postgres, that's considered "willful" instead of "innocent"
> infringement and the penalties are different.  Its effect on our project
> is the same, though: we have to rip out the code in a hurry.
> 
> Maybe we should just relicense PostgreSQL as Apache and cover all of the
> patent issues ;-)

LOL, but the existing Apache 2.0 license doesn't extend the patent grant
to derived works, including closed source ones, so that would not seem to
help.  If it did extend to all derived works, the patents would be
unenforceable.

> >>> Of course, never mention known-patented ideas in any community forum,
> >>> including this email list.
> >>
> >> I just found out that Citus Data has patent applications pending, so
> >> viewing Citus Data source code has the same problems as Greenplum.
> > 
> > Actually, it might only be their closed source software that contains
> > patents, i.e. not pg_shard.  I will check and report back when I can
> > unless someone else reports here first.
> 
> I will ask Citus Data for an official statement.  I will point out that
> cstore_fdw is Apache-licensed, which also contains a patent grant.

Good point.  So we have CitusDB, cstore_fdw, and pg_shared, the later
two are open source.  Ideally only the closed source CitusDB implements
patented ideas they own.

Let me add that this is more than hypothetical.  While we don't think
any of these companies would sue the community for patent infringement,
they could sue users, and the company could be bought by a sinister
company that could enforce those patents.  For example, few had problems
with Sun's control over Java, but when Oracle bought Sun, more people
were concerned.  Someone could buy the company _just_ to sue for patent
infringement --- happens all the time.

--  Bruce Momjian  <bruce@momjian.us>        http://momjian.us EnterpriseDB
http://enterprisedb.com

+ As you are, so once was I. As I am, so you will be. +
+ Roman grave inscription                             +

On 1 November 2015 at 07:47, Bruce Momjian <bruce@momjian.us> wrote:

On Sun, Nov  1, 2015 at 01:27:13AM -0500, Bruce Momjian wrote:
> On Fri, Oct 30, 2015 at 04:47:35AM -0400, Bruce Momjian wrote:
> > Therefore, I caution people from viewing the Greenplum source code as
> > you might see patented ideas that could be later implemented in
> > Postgres, opening Postgres up to increased patent violation problems.  I
> > am also concerned about existing community members who work for
> > Pivotal/Greenplum and therefore are required to view the patented source
> > code.  The license issue might eventually be improved by
> > Pivotal/Greenplum, but, for now, I think caution is necessary.
> >
> > Of course, never mention known-patented ideas in any community forum,
> > including this email list.
>
> I just found out that Citus Data has patent applications pending, so
> viewing Citus Data source code has the same problems as Greenplum.

Actually, it might only be their closed source software that contains
patents, i.e. not pg_shard.  I will check and report back when I can
unless someone else reports here first.

While you are doing that, please also check EnterpriseDB. My information is that there are patents filed there, so we must check that just as much as any other company or person. If you didn't know before, you do now.

I am disappointed that your approach to this appears unbalanced and partisan. Worse, Greenplum have been quite vocal about their intentions, so any feedback you have could easily have been given many months ago, not on the day of their announcement. I think you should have declared this situation in a very different way to the way you have approached this. 5 minutes thought on whether other companies might also have been affected would have been sensible, plus the whole thing could have been discussed completely offlist. If you do discuss things on-list then you should at least state for the record that you are an EnterpriseDB employee when discussing your concerns, since that is likely to have a material affect on how this situation is viewed by anyone worried by your post.

On 11/01/2015 06:37 PM, Bruce Momjian wrote:
> Let me add that this is more than hypothetical.  While we don't think
> any of these companies would sue the community for patent infringement,
> they could sue users, and the company could be bought by a sinister
> company that could enforce those patents.  For example, few had problems
> with Sun's control over Java, 

You only say this because you're not part of the Java world.  LOTS of
people had issues with Sun's control over Java; some of them even went
to court.

> but when Oracle bought Sun, more people
> were concerned.  Someone could buy the company _just_ to sue for patent
> infringement --- happens all the time.

Not as often as you'd think, and it hasn't happened in the database
world yet, for some good reasons.  This is all besides the point,
though; PostgreSQL has been accepting contributions from patent-holding
companies for over a decade, and that doesn't seem likely to stop any
time soon.  Greenplum is not in any way special, especially since we
already accepted contributions from Greenplum Inc. back in 2005-2006.

Overall, this thread seems designed to kick up a lot of fuss with no
potential useful outcome.  How about we terminate it now?

-- 
Josh Berkus
PostgreSQL Experts Inc.
http://pgexperts.com

On Mon, Nov  2, 2015 at 10:36:48AM -0800, Josh Berkus wrote:
> > but when Oracle bought Sun, more people
> > were concerned.  Someone could buy the company _just_ to sue for patent
> > infringement --- happens all the time.
> 
> Not as often as you'd think, and it hasn't happened in the database
> world yet, for some good reasons.  This is all besides the point,
> though; PostgreSQL has been accepting contributions from patent-holding
> companies for over a decade, and that doesn't seem likely to stop any
> time soon.  Greenplum is not in any way special, especially since we
> already accepted contributions from Greenplum Inc. back in 2005-2006.
> 
> Overall, this thread seems designed to kick up a lot of fuss with no
> potential useful outcome.  How about we terminate it now?

I am posting this at the request of Josh Berkus, who wanted
clarification on some issues.  FYI, I have been speaking in this thread
as a community member, and not as a member of core, and made some
mistakes in my handling of this --- my apologies.

First, I have always seen the best intentions from every patent holder I
have worked with in relation to Postgres, including the many Pivotal
employees I have talked to.  I understand the value of those patents to
their companies and their companies' valuation, and releasing any kind
of rights to held patents is never an easy decision.  I know all
companies involved are trying to deal with this complex issue in the
best possible way.

Second, if Pivotal had to chose one license to release their code under,
Apache 2.0 was absolutely the best one, because of the patent clause.  I
will continue to work with them or anyone else as requested to see if
there is an even better approach.

The crux of my concern is that a patent in close-source software is
barely visible --- it might be mentioned in marketing material or
documentation, but that is unlikely.  When something is released as open
source, by definition, the patented idea is visible in that code.  In a
strange twist of fate, open source actually allows more chances for
seeing patented ideas than closed source.  I should have made this clear
in my initial post, and there is nothing Greenplum-specific about any of
this.  My lack of clarity on this caused much confusion --- again my
apologies.

Steven might be right that if you don't know the patent is there, you
are not any more culpable than if you discovered the technique on your
own.  However, the odds of you getting a patented idea from looking at
the source is probably higher than the odds of you coming up with the
idea on your own.

Anyway, I just wanted people to be aware of these risks.  I doubt we can
really do anymore than warn folks as there is just no good boundary on
how to avoid problems, as Andres and Simon pointed out.  Josh is right
that patent problems have been a rarity, and I have every hope that this
will continue.

--  Bruce Momjian  <bruce@momjian.us>        http://momjian.us EnterpriseDB
http://enterprisedb.com

+ As you are, so once was I. As I am, so you will be. +
+ Roman grave inscription                             +