Hi All, On the GMP website we find that MPIR "have also removed the author name from our 140 page manual".
Roughly speaking, whilst not strictly correct as written (we haven't "removed" anything), my reading of the license suggests there *is* an issue that needs to be sorted out here. As of MPIR 2.0.0 we switched over to using the GFDL version 1.3 in the MPIR manual, so that it is possible to include sections from the more recent GMP manual in the MPIR manual. One of the conditions of the GFDL 1.3 is apparently that the author's name be given on the title page. At present we give no author's name at all and never have! However current GMP documentation has "Torbjorn Granlund and the GMP Development Team" listed as authors. Thanks to the GMP developers for pointing this out. However, there is an accredited technique for doing so, it involves a short email which goes like this, "Dear MPIR Developers, We have just noticed that currently the MPIR manual does not have the Author names on the Title page as required by the GFDL version 1.3. Please note the authors listed in the current GMP manual on its Title page. We would appreciate it if you could make the necessary adjustments to your manual. Regards, The GMP Team." Instead we find out about this issue on the GMP website! I can promise that if a short polite email such as this is sent to the MPIR team in future, the issue will receive our fullest attention and will receive a similarly polite response notifying the GMP Team that we have made the adjustments requested. Please be specific about any issues you might note. Generalised accusations with no detail, which later prove to be false, as has happened repeatedly in the past, are not welcome! Also, threatening legal language is not required in such correspondence. We are a fellow Open Source project and not a large company with a proprietary license and big legal department! It would also be preferred if this correspondence was sent to a member of the MPIR Development Team instead of the leader of the Sage project, who is not an MPIR devel (and infinitely busy). Although our manual came to being as a modified version of a GFDL 1.2 manual, it does now include a few small bits and pieces from the new manual and is therefore now GFDL version 1.3 (yes we remembered to include the license for that!!). To reflect this, I believe we should change the title page to reflect the authors of the manual. I propose we add, "Original Authors: Torbjorn Granlund and the GMP Development Team" followed by "Subsequent modifications: William Hart and the MPIR Team" (and of course anyone else who modifies the manual in a significant way from now on, e.g. J. Moxham). Please refer to the statement that says that up to five authors *MUST* be listed, if they exist. I wonder if anyone else modified the GMP documentation? Who else is included in the "GMP Development Team" that should be listed as an author. Or is the document entirely written by one individual? Note the requirement to have this information on the title page seems to be a new one with the GFDL v1.3, or at least no such information was in GFDL v 1.2 documentation. Note that the original defamatory information posted on the GMP website about MPIR has been removed. Thanks to the GMP Team for removing this! As mentioned before, the FSF conducted an audit of MPIR some time ago, and after correction of a couple of minor issues (mentioned in an earlier post) was happy to let us know that they consider us to be in compliance. Certainly none of the original claims made on the GMP website or in private correspondence that we had been caught "red-handed" stealing code from recent GMP and downgraded the license headers, were supported by that audit. Now to the other current inaccuracies: 1) "a renamed GMP" - actually MPIR contains a vast quantity of new material not contained in GMP. See the MPIR website and past release announcements for details. It is incorrect to imply that we just renamed GMP to MPIR. The correct term is "fork", not "rename". 2) "initially based on GMP 4.1.3" - there may be some truth to this in that the Sage version of GMP which contained some patches, was quite possibly based on GMP 4.1.3. for a while. I haven't bothered checking if this is correct. However, the MPIR project came to life later and we deliberately discarded the Sage spkg and forked directly from GMP 4.2.1. So MPIR is not a fork of GMP 4.1.3. I don't know what to make of the note, "although they inadequately for a long time released their GMP version under LGPL 2". Perhaps they are referring to the fact that for a long time we said on our website that MPIR was LGPL v2+ (an abbreviation for LGPL v2.1+). As far as I know, the library itself was correctly licensed though. In short, either they are confused, or they are objecting to the fact that we wrote LGPL v2+ instead of LGPL v2.1+ on our website, which we feel is a bit precious. Either way, it doesn't seem to be a problem now. 3) "Later versions are copied from GMP 5.0, and also correctly uses LGPL 3" - strictly incorrect. Whilst we don't deny that we *do* use the correct license, MPIR 2.0.0 is not a copy of GMP 5.0.0. It uses _some_ of the recent code that has been added to GMP 5.0.0 (though modified for inclusion in MPIR). We have been very selective in what we have included from GMP 5.0.0. This is legal and perfectly legitimate. Is it intended to appear to be a negative thing to have used some GMP source code in another open source project? Well, what are the implications of that? Do we not give copious credit for this code on our website and in our documentation? Is there something academically immoral here? Illegal? OK, let's just presume they are proud of their code and chuffed that we make use of some of it and want to tell the world about that fact. 4) "It is maintained by a group of people with funding from Microsoft." - if you count a donation of $5000 from MSR (a long time ago) - about half of which was used to support MPIR development and some free copies of their development software to be "funding from Microsoft", then this is a correct statement. The implication that we currently have cash funding from Microsoft and that we use it to support our development is one that I would very much like to be true and would wear as a badge of pride were it so. I am perfectly happy for the rumour to be circulated that we do currently receive such funding. N.B: one of our developers *does* sometimes get free copies of some MS development tools and in fact will do so shortly to help with maintaining our imminent release of an MSVC 2010 version of MPIR! We are *very* thankful to MSR for this continued support!! What a surprise that MPIR is now LGPL v3+, MSR don't like this license and they still support us! So the former claims that MS demanded we "change" the license to v2+ as part of being "bought out" by MS were incorrect after all. What a shock! 5) "They bootstrapped their project with a defamation campaign against GMP" - the definition of defamation requires that public statements be made which are false. Some of the statements were statements of fact that were a matter of public record (e.g. the lack of a public repo - since rectified by GMP, or the license changed before adequate consultation), or the cold attitude towards Windows development (nothing makes us believe this doesn't continue today and roughly speaking it appears to be a matter of policy for GNU projects and their resources). Other statements weren't even worded as statements, but were posed as questions. However, let me say this. I personally *do* regret making many of those statements (regardless of their veracity). I *apologise unreservedly* for upsetting the GMP developers. This flame war, whether well-intentioned or not, is much to be regretted. It would have been better to simply state we forked because we wished to support Windows development, have an open repository and, at the time, continue with an LGPL v2.1+ licensed library whilst GMP moved on with v3+, so that other v2+ projects had time to assess v3+. That would have been more reasonable on our part. We were naive in the extreme to not expect the hostile response we got. This we regret deeply. 6) "replaced every string "GMP" with "MPIR"" - this was actually reviewed by an academic who stated that they felt this had been done very carefully. Actually some "GMP"'s remain, where applicable. I fail to see the problem with this. Perhaps the GMP developers care to state their objections more clearly. They do realise that MPIR is a fork of GMP not simply a rename, right? Or perhaps they are not prepared to accept that a significant quantity of bignum expertise exists outside the official GMP project, and always has. Maybe they think that all we have been capable of is renaming their code and calling it a fork!? It's amazing the speedups we've gotten over the GMP project, for years, from this "rename". Can someone "rename" GMP 6 already! 7) "and released this manual as their own original work" - copyright notices are preserved, so the recent "title page" thing, notwithstanding, I reject this statement. The remainder is just FUD and doesn't merit a response. Please note, a number of bugs in GMP releases have been patched in MPIR. I am very confident of the overall quality of MPIR. However note that, as always, MPIR comes with NO WARRANTY WHATSOEVER, etc., etc., same as GMP. There *ARE* bugs in MPIR, just the same as there *ARE* bugs in GMP. If you believe the result of any computation from a library this complex, or mathematical software in general, you have rocks in your head (to put it bluntly). Please note, my involvement with the MPIR project at this point is only to help in the transition to new leadership. I am not actively involved in contributing code to MPIR at any significant scale and have stepped down as leader of the project. The project continues to thrive and a new release will be made very shortly. I know there are people eagerly waiting for this release due to our support of MSVC 2010 through the extremely hard work of Dr. Brian Gladman over the past months! I wish the MPIR team all the best in future and hope for a new era of cooperation with the official GNU project. In particular I hope that the documentation issue noted above can be resolved as soon as possible and the GMP team notified. I believe this affects MPIR 2.0.0 and following only (please check the old documentation license to confirm if this is the case). Bill. -- To post to this group, send an email to sage-devel@googlegroups.com To unsubscribe from this group, send an email to sage-devel+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
