GMP license problem, anyone?

Bill Hart goodwillhart at
Thu May 29 21:37:25 CEST 2008

>I've read and reread your long posts, but I have not understood the
>reasons for your stated problems with v3+.

>Have you read the v3 license text, and have a specific problem with
>some part of it (which?), or do you have principal objections to the
>fact that the license was updated?  In the latter case, one would
>solemnly expect you to be using GPL v1.0.  ;-)

Firstly, Linus Torvalds and other linux kernel hackers didn't want v3.
Linus said the kernel would remain v2 and I quote: "Conversion isn't
going to happen."

Secondly, Microsoft Research are one of SAGE's customers. Microsoft
has a policy of not even running software with v3.

Thirdly, the licenses of all the major components of SAGE that are GPL
licensed did not change to v3+ as GMP did. They either changed from v2
only to v2+ or they remained v2+ as they have always been.

>From my point of view, where I sit, I would need a reason to change to
v3+. No one has given me a specific reason why I should.

The principle objection that I have to v3 is that it's not v2. I like
version 2 and license my code under v2+. I didn't need all the extra
stuff in version 3, there was no pressing need for it, I didn't care
for it when it came along. It just adds a whole load of extra stuff
which I don't care about. In other words I feel v3 is overreaching for
a general purpose software license.

>  I should also point out that the work we are doing on our own version
> of GMP (which I will call X to save typing) and the work I do on
> FLINT, is in many respects insignificant in comparison to the much
> larger businesses which you deal with.

>Eh, what businesses?

I was under the impression that AMD had paid you in the past and that
Mathematica currently are supporting the development of GMP. It's good
to have misconceptions like that dealt with.

>  At any rate, my point is that the type of sponsorship we are
>  attracting (and targetting) is likely to be very different to the sort
>  of sponsorship that GMP gets,

>If you have managed to get sponsorship for said fork, you're more
>financially successful than "the official GMP project" (to use your
>words).  The total amount of funding for the last 8 years amounts to
>15000 USD.  :-)

That's about 65000 USD short of the amount I was told someone was
requested for merging of a patch by yourself into GMP. Well apparently
they didn't pay you. Again, good to have misconceptions dealt with.
There are some widespread misconceptions out there, and one of them is
that you ask very high fees for merging patches.

>  Anyhow, now that you see what we are doing, back to the licensing
> issue. Clearly FLINT was badly hurt by the license change. It means I
>  can't use GMP v4.2.2 or above for FLINT.

>Please explain why you "can't" use it.  This is the essence of my

Because I wish FLINT as a project to be covered by v2 of the license.
If someone modifies it and distributes their changes under v3, no
problem, at least people can come back to the FLINT project and get
the version I wrote with v2+ of the license. If I start using code
from GMP that is covered by v3+ and not v2+ then I can't have overall
v2+ on my package.

>  But is there realistically anything you can do about this? Now that
>  you have turned your copyright over the the FSF, can you change the
>  license on that code? You don't have control over the licensing any
>  more.

>The FSF, that's us Free Software hackers.

You make it sound like you own your copyright. You didn't answer my
question. Could you change the license back, even if you wanted to?

My question to you is, why did GMP change to v3+ and not remain with
v2+. What extra freedoms does it give users of your code or what extra
protections does it give you and how is that relevant specifically to

What reason would I have for wanting to switch to v3+ on any of my
code or any code I include.

People say, "v3 must exist for a reason, so the right thing to do is
switch". Well what is that reason?

As to insinuations and misconceptions. Yes, it looks like there are
some misconceptions. But there are some other reasons for forking GMP
which I didn't give, as it is not to do with licensing.

Please correct me if I have a misconception or am merely making an insinuation:

1) GMP has a slow release cycle. Patches are only considered for
inclusion in the next major release of GMP, and that next major
release is 5, which numerous places on your website you say is years
off. This means we won't see decent assembly support for Core 2 or AMD
64 for years.

2) So far the GMP project has refused to accept patches which will
allow GMP to build under [insert proprietry OS and compiler here].

3) The development process for GMP is not transparent. What gets in is
solely the maintainers decision.

OK, I see now that William Stein has already given a list of reasons
to the list, so I don't need to give them all here from my point of
view. He's been much more direct than I was ever going to be.

My main concern here was the licensing issue anyway. I only mention
the above because you wanted to know some of the other reasons for the
fork. So I'll stop here.

Bill Hart.

More information about the gmp-discuss mailing list