witam,
Michał wrote:
Boud Roukema wrote:
what is the purpose of GPL? creating better, community driven software. imho it is more than a license, it is a philosophy. your "software" just
It is not philosophy. It is politics, because it relates to the power that some people have over other people, and the power that people do *not* have over other people.
The intention is to minimise the power that people have over others to the minimum necessary - in the context of computer code.
It depends what you mean by "politics" and "philosophy" I believe. in fact it is both. but developers seem to prefer the term "philosophy" since it is more positive in meaning.
i guess this may be an anglophone/francophone difference. Personally i prefer precision rather than political correctness.
do not fit GPL. it is a set of mathematical operations. how the hell do you want people to use it/trace bugs/improve when you only provide hardly readable code??????
we could argue about the quality of the code but I do not think will do any good.
The point is not to argue about the quality of the code.
The whole point of the GNU GPL is that concrete discussion of the code can happen according to rules which encourage improvement of the code.
Without distribution under GNU GPL, it is difficult to get the *feedback* that someone has the opinion that the code is "hardly readable".
I agree.
:)
but without providing high-quality code and docs it is hard to get useful feedback.
It's sufficient to improve by the same fraction.
Consider a code of low initial quality p(0) = 0.3 with low docs quality q(0)=0.3, versus code with high initial quality p'(0)= 1.0 and high initial docs quality q'(0) = 1.0. In the first case there is on average one iteration of feedback each week which yields moderate improvement in the combined quality x = pq of 10% (alpha = 1.1), in the second case the initial version is only distributed after one year, and then gets feedback yielding improvements of 10%.
p(0)=q(0)=0.3 p'(0)=q'(0)=1.0
x(t) = p(0)q(0) alpha^n (n= weeks) x'(t) = p'(0)q'(0) alpha^(max(0,n-52))
which gives for t \ge 52:
x(t)= 12.8 alpha^(n-52) x'(t)= 1.0 alpha^(n-52)
Conclusion: many iterations through a positive feedback loop starting from an imperfect beginning can give a better result than waiting a long time for a "perfect" beginning to be ready.
On the other hand, if there is a delay of between several days and a few weeks for having the right to use the code, and then a further delay of unknown length to redistribute modified versions, the cooperation between X, Z, W and U is unlikely to happen, especially given that Y feels unhappy with the code and may influence the delays.
I can bet that Kriss would be more that happy to accept any help with porting his software to other languages. Have you tried to chat with him about licence issues?
If you wished to be a neutral, third party moderator to discuss with a certain person regarding certain software, for a publicly archived discussion (like this one) regarding porting of cosmologically useful software to a version distributed under the GNU GPL or a compatible licence, and if you convinced that person and other people interested that you would be a good moderator (meaning that you clarify the discussion, stop different people from flaming or threatening each other, ask them to apologise or find other ways of getting people to cool off if they start flaming, etc.), i would be happily accept this.
My prediction is that this would be not be accepted by some of the people, but i would be glad to be wrong.
As I still am quite sure about there is no licence that suits scientific software and gpl is not the best. that is why some people refuse gpl-ing. I myself would be very happy having a nicer licence for sci soft.
Personally i'm not convinced of the need. I'm aware that if you spend a lot of time on software and don't write many publications, you need to get credit for the software. The GNU GPL already guarantees that the original author's name remains, and IMHO the cosmo community is small enough that people writing, or more importantly, maintaining good software do get credit. Just because it's not formally part of these strange formulas attempting to measure scientific productivity in PL, does not mean the credit is not there.
However, you want a modified licence, so please have a look at:
http://www.gnu.org/licenses/gpl-faq.html#ModifyGPL
What do think of a licence which is identical to the GNU GPL, but we remove the preamble (unless FSF agrees that we may include it in our modified version after discussion), we modify the end-wording, and we add a term/condition like:
SSL-13.0 - proposal for Scientific Software Licence-term/condition 13.0 ---------------------------------------------------------------------- 13. Any publication in a recognised scientific journal which makes use of this software must contain the acknowledgment: "Use has been made of PROGRAM-NAME which can be obtained at PROGRAM-URI", where PROGRAM-NAME is the name of the Program and PROGRAM-URI is a Universal Resource Identity showing where the full machine-readable source code of the Program can be downloaded from." ----------------------------------------------------------------------
and then write to licensing at gnu.org to ask for advice.
My guess is they would only complain about the compatibility question:
http://www.gnu.org/licenses/gpl-faq.html#WhatIsCompatible
In which case i would suggest in addition:
SSL-2.1 - proposal for Scientific Software Licence-term/condition 2.1 ---------------------------------------------------------------------- identical with GNU GPL v2 term/condition 2. except for the following:
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% "But when ... wrote it. " %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
becomes
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% "But when you distribute the same sections as part of a whole which is a work based on the Program, the distribution of the whole must be on the terms of either this License, whose permissions for other licensees extend to the entire whole, and thus to each and every part regardless of who wrote it, or of the GNU General Public License, as published by the Free Software Foundation (either version 2 of the License, or (at your option) any later version) whose permissions for other licensees extend to the entire whole, and thus to each and every part regardless of who wrote it.
If you apply the GNU General Public License rather than this License to the whole, then you are requested (but not required) to include a file called README.acknowledgments in the machine-readable source code which includes the text of term/condition 13. below, with the phrase
"Any publication ... acknowledgment:"
replaced by
"It is kindly requested (but not legally required) that any publication in a recognised scientific journal which makes use of this software should contain the acknowledgment:" %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
----------------------------------------------------------------------
In short, this means that if an SSL protected package is combined with a non-GNU GPL package, then:
- if the other package is also under SSL, the new one can also be SSL - if the other package is GNU GPL, then the new one *must* be GNU GPL (condition 2 of GNU GPL), but the user is *requested* to add a file converting term/condition 13 into a README.acknowledgments file which *requests* the user to write an acknowledgment - if the other package is under some other licence, it may or may not be possible to distribute the combined package, it depends on that other licence.
IMHO, this is as strong as we could get while retaining GNU GPL compatibility. If we try to make the request into a requirement, then my worry is it would be GPL incompatible.
In fact, compatibility is one of the big issues of the GNU GPL. It enables modules to be combined together in ways far beyond what the original authors ever could do alone. Remember: there are 6 billion of us. That's a lot.
Anway, you wanted a licence: what do you think of this one?
If we go through a few iterations, with a bit of feedback :), then i'd be happy to write to licensing at gnu.org, point them to our discussion, and see what they think.
A simpler alternative to SSL-13.0 + SSL-2.1 would be:
SSL-13.1 - proposal for Scientific Software Licence-term/condition 13.1 ---------------------------------------------------------------------- 13. It is kindly requested (but not legally required) that any publication in a recognised scientific journal which makes use of this software should contain the acknowledgment: "Use has been made of PROGRAM-NAME which can be obtained at PROGRAM-URI", where PROGRAM-NAME is the name of the Program and PROGRAM-URI is a Universal Resource Identity showing where the full machine-readable source code of the Program can be downloaded from."
If you distribute this Program as part of a whole which is a work based on the Program, and you apply the GNU General Public License (GNU GPL) to that whole, then you are kindly requested to copy this term/condition 13. to a file called README.acknowledgments (if it is not already present) and include it in the machine-readable source code. Since further distribution of the whole would occur under the GNU GPL, there would be no further obligation to include the file, only a kind request. ----------------------------------------------------------------------
pozdr boud