February articles on scientificcomputing.blogspot.com
Scientific Computing
http://scientificcomputing.blogspot.com/2006/02/security-risks-in-web-based-scientific.html
Maple handwriting recognition: useful or gimmick?
http://scientificcomputing.blogspot.com/2006/02/maple-handwriting-recognition-useful.html
Open source technical software - dead end?
http://scientificcomputing.blogspot.com/2006/02/open-source-technical-software-dead.html
Mathematica released for Mac Intel
http://scientificcomputing.blogspot.com/2006/02/mathematica-released-for-mac-intel.html
Firefox update Released
http://scientificcomputing.blogspot.com/2006/02/firefox-1501-released.html
MAA Placement tests in MapleTA
http://scientificcomputing.blogspot.com/2006/02/maa-placement-tests-in-mapleta.html
Statistical Inference package for Mathematica
http://scientificcomputing.blogspot.com/2006/02/statistical-inference-package-for.html
Rumored product from Maplesoft: Blockbuilder for Simulink
http://scientificcomputing.blogspot.com/2006/02/rumored-product-from-maplesoft.html
" >
> Open source technical software - dead end?
> http://scientificcomputing.blogspot.com/2006/02/open-source-technical-software-dead.html
>
Interesting points but I don't share your conclusions.
First I don't believe that any future important open source software
should probably be in the (Axiom, Maxima, Reduce, Scilab) list.
The analogy argument with non scientific softwares applied
in 1990 would have missed Linux.
Then your arguments 1 and 2 also apply for commercial scientific
software. The ratio of people who can contribute to open-source
scientific software is smaller than for general software, but there
is no reason that the percentage of open-source vs closed-source
should also be smaller.
The question of financial backing is more complex than your
argument 3, because many open-source scientific software
development is made by academics who are paied.
For argument 4, the learning curve should be a major concern
for any open-source developpers, it is true that it was
underestimated, it is no longer the case for general open-source
software nowadays, so it's probably only a question of time before it
is no longer the case in the scientific open-source softwares.
For argument 5, I think that your comparison is biaised since
you take examples of commercial softwares who are (were) in
a monopoly position, which is not the case for commercial scientific
softwares.
Long live to open-source scientific software!
Richard J. Fateman
none wrote:
> Scientific Computing wrote:
>
>> Open source technical software - dead end?
>> http://scientificcomputing.blogspot.com/2006/02/open-source-technical-software-dead.html
>>
>>
>
> Interesting points but I don't share your conclusions.
> First I don't believe that any future important open source software
> should probably be in the (Axiom, Maxima, Reduce, Scilab) list.
> The analogy argument with non scientific softwares applied
> in 1990 would have missed Linux.
The analogy, that yesterday's commercial program becomes today's
open source DOES fit. ATT UNIX "became" in some sense, BSD UNIX,
and Linux. [not the source code but the concept.]
> Then your arguments 1 and 2 also apply for commercial scientific
> software. The ratio of people who can contribute to open-source
> scientific software is smaller than for general software, but there
> is no reason that the percentage of open-source vs closed-source
> should also be smaller.
I don't understand this. It looks to me like the number of people
contributing to the (open source) Maxima is larger than the company
(Macsyma Inc) had working on the program, at least at some times
in its evolution. Current commercial CAS spend substantial programmer
time on rather shallow non-mathematical tasks such as revising the
GUI menus, adding handwriting recognition, etc.
Finding talented people (programming + math) is hard, but paid or
free may not be the factor that matters most.
Open-source is not free like free beer, so it is easy (and correct)
to attack this concept as having costs for use. Who should
pay for free computer algebra system development then?
Who knows. Maybe a rich philanthropist. Maybe some company that
wants to use it. In which case those who insist on including
a GPL "poison pill" are limiting the funding possibilities.
DP
Richard J. Fateman wrote:
>
>
...
> Open-source is not free like free beer, so it is easy (and correct)
> to attack this concept as having costs for use. Who should
> pay for free computer algebra system development then?
>
...
> In which case those who insist on including
> a GPL "poison pill" are limiting the funding possibilities.
>
This argument has been propagated by those having commercial interest
to include free code (mostly with BSD license) into commercial codes
without contributing back anything (freedom to reap profit).
Software developers wanting that their donated code remains free insist
on using GPL, this is why most of the open source code (in sourceforge.org
for example) is licensed with the GPL.
The "poison pill" image smells lack of a better argument.
It has been used, as well as "cancer", to characterize OSS in FUD campaigns
or articles payed by Microsoft. Microsoft dislikes extremely the GPL,
and spends 100's of M$ in FUD campaigns instead of improving its products.
To me it means the GPL is a very effective competition.
In conclusion perhaps the GPL is limiting funding by commercial sources,
but surely the GPL is encouraging programmers to develop OSS.
Richard Fateman
"DP" <pfen...@obs.unige.ch> wrote in message
news:44096CF3...@obs.unige.ch...
... snip...
>
> In conclusion perhaps the GPL is limiting funding by commercial sources,
So, after arguing that GPL is so great, you end up agreeing with me.
> but surely the GPL is encouraging programmers to develop OSS.
It is also encouraging them not to quit their (paying) jobs delivering
pizzas.
RJF
>
DP
Richard Fateman wrote:
> "DP" <pfen...@obs.unige.ch> wrote in message
> news:44096CF3...@obs.unige.ch...
> ... snip...
>> In conclusion perhaps the GPL is limiting funding by commercial sources,
>
> So, after arguing that GPL is so great, you end up agreeing with me.
No.
I just wanted to suggest that commercial funding is perhaps not always
necessarily the only and best source for software development.
Commercial software has often characteristics that users dislike (horrible
license conditions, slow bug correction, lack of flexibility, high cost).
A lot of software in use, in particular CAS, has been developed in
universities or state financed institutions. It is not hard to see
that for general usage software it would make sense economically,
like for public services such as the road system, to increase the
scope of government supported free software. Specialized software
will always exist so there will always be a market for commercial software.
>> but surely the GPL is encouraging programmers to develop OSS.
>
> It is also encouraging them not to quit their (paying) jobs delivering
> pizzas.
This caricature of free software programmers is unfair.
While some are fully payed to produce free software (Linus Torvald)
others do it for altruistic motives, or for fun. Like scientists, or artists.
There is in fact a large diversity of conditions allowing individuals
to produce free software.
Dan
Richard Fateman
send me a valid email address.
"DP" <pfen...@obs.unige.ch> wrote in message
news:4409F36F...@obs.unige.ch...
<snip>
> I just wanted to suggest that commercial funding is perhaps not always
> necessarily the only and best source for software development.
So we agree again. I pointed out that GPL inhibited certain kinds of
funding, namely from (some) commercial sources.
> Commercial software has often characteristics that users dislike (horrible
> license conditions, slow bug correction, lack of flexibility, high cost).
Whereas non-commercial software has characteristics like high cost
(you must hire a programmer to spend a day figuring out why it doesn't
install on your computer), lack of flexibility (oh, I didn't put in that
feature because if you really need it you can put it in yourself and
recompile), slow bug
correction (sorry, I was on vacation, and besides, I have to deliver
this pizza, do you want to hire me?), horrible license conditions (you
are forbidden from just giving someone the binary; you have to display all
this legal
stuff etc. and then tell them they better get a lawyer if they plan to use
it in some ways.).
>
> A lot of software in use, in particular CAS, has been developed in
> universities or state financed institutions.
But the most popular ones are commercial. Why do you suppose
Mathematica, Maple went that way? And Mupad too?
> It is not hard to see
> that for general usage software it would make sense economically,
> like for public services such as the road system, to increase the
> scope of government supported free software.
Apparently the government of the US doesn't agree. Otherwise it
would not be suppressing the distribution of existing IRS software to fill
out the government's tax forms. It would compete with commercial
software...
....
> There is in fact a large diversity of conditions allowing individuals
> to produce free software.
Mostly having another job that pays you. Like being a professor
or a grad student. How many people do you know who are
paid specifically to produce / improve / fix bugs / in a
GPL computer algebra system?
I repeat, I am not against giving away software. I do it
quite frequently.
RJF
>
> Dan
Jay Belanger
"Richard Fateman" <fat...@cs.berkeley.edu> writes:
> Since this is largely off topic and argumentative, I invite DP to
> send me a valid email address.
Is the licensing of symbolic math software off-topic here?
(I really don't know.)
> "DP" <pfen...@obs.unige.ch> wrote in message
> news:4409F36F...@obs.unige.ch...
...
>> I just wanted to suggest that commercial funding is perhaps not always
>> necessarily the only and best source for software development.
>
> So we agree again. I pointed out that GPL inhibited certain kinds of
> funding, namely from (some) commercial sources.
That's not agreeing.
>> Commercial software has often characteristics that users dislike (horrible
>> license conditions, slow bug correction, lack of flexibility, high cost).
>
> Whereas non-commercial software has characteristics like high cost
> (you must hire a programmer to spend a day figuring out why it doesn't
> install on your computer),
Unless it installs easily on the computer.
I once spent a week trying to install Mathematica on my computer,
using the support that my university payed through the nose for.
I eventually decided it wasn't worth it, and stuck to free software.
I rarely have trouble, and talking to the people here who are in
charge of installing software, they seem to have more trouble with the
proprietary software than free software. (Just my impression from
conversations; I never asked them specifically.)
> lack of flexibility (oh, I didn't put in that feature because if you
> really need it you can put it in yourself and recompile), slow bug
> correction (sorry, I was on vacation, and besides, I have to deliver
> this pizza, do you want to hire me?),
In my experience, bug correction in free software is faster than
proprietary. At any rate, do you have any examples of the above in
mind, or are you just making things up? If you do have an example of
the above, it was bad software and should be avoided. The license
really wasn't the issue.
>> It is not hard to see
>> that for general usage software it would make sense economically,
>> like for public services such as the road system, to increase the
>> scope of government supported free software.
>
> Apparently the government of the US doesn't agree.
What the US government does and what would make sense aren't
necessarily the same thing.
Jay
Gabriel Dos Reis
[...]
| >> Commercial software has often characteristics that users dislike (horrible
| >> license conditions, slow bug correction, lack of flexibility, high cost).
| >
| > Whereas non-commercial software has characteristics like high cost
| > (you must hire a programmer to spend a day figuring out why it doesn't
| > install on your computer),
|
| Unless it installs easily on the computer.
How often?
-- Gaby
Richard Fateman
"Jay Belanger" <bela...@truman.edu> wrote in message
news:873bhxl...@vh213602.truman.edu...
>
> Is the licensing of symbolic math software off-topic here?
> (I really don't know.)
>
I think a discussion GPL vs BSD vs commercial vs. ... is about software
in general, not CAS.
There is no moderator here, so if you feel CAS are in some important way
different from
other software in their relation to licensing, I guess it is on topic.
sci.math.symbolic seems to
suffer from OT postings, though there are worse newsgroups in that respect,
and only the
occasional troll.
RJF
DP
Dr. Fateman,
I think you are responsible to have started unworthy attack against
the GPL and caricature of free software programmers. If you spread
FUD in public you must expect public reactions.
To come back on topic, as along contributor to Macsyma you should be
well positioned to understand some major drawbacks of commercial CAS.
Macsyma is a prime example of wasted efforts having deserved the field,
that would have been avoided with the GPL.
Dan
Dave (from the UK)
> Open source technical software - dead end?
> http://scientificcomputing.blogspot.com/2006/02/open-source-technical-software-dead.html
You say: "When you look around at large scale open source projects they
are nearly all Commercial software that has been released. e.g. Firefox
was once Netscape. OpenOffice was once StarOffice."
So how about Linux then? Without too much doubt the most successful
open-source project about.
Whilst I realise this a computer algebra newsgroup, why does the link
above, where the title has nothing about computer algebra, pick on 3 CAS
systems when it says:
"On this basis one can probably assume that any future major scientific
software open source project already exists. It is either one of the
existing open source projects (Axiom, Maxima, Reduce, Scilab) or it is
one of the commercial products if its owners fail to make it pay or fail
as companies (Matlab, Mathematica, Maple, Mupad etc)."
"So let us turn our attention to the three main large open source
projects Axiom, Maxima and Scilab."
That is a pretty narrow definition of scientific software.
Most of your other points I tend to agree with.
1) I agree the knowledge required to write scientific software is high.
This tends to reduce the number of contributors. I wrote some software
for solving a transmission line problem
but its capabilities are a small subset of expensive commercial
software. But to write much more capable software you need a large bunch
of PhDs putting a lot of effort/time into it. I somewhat doubt it is
practical for that to happen. Hence for tasks like that, which really
need a huge amount of knowledge in the area concerned, it is not very
likely to happen in open-source.
2) You say "Large technical computing systems have a lot of internal
dependencies."
I'm not convinced that should be any more/less of a problem than with
other software.
3) You say "To support the costs, free software needs major financial
contribution."
I am not convinced that is true. There are some very popular software
projects that don't have huge financial backers.
4) "The major selling point of free software is "it's free". But of
course there are other costs to installing new software. Principally the
training cost."
I would agree with that.
To me, the biggest "problem" is that scientific software is often
obtained for nearly zero cost in a university. I had access to
Mathematica, Matab, Labview... whatever else I wanted really. As a
member of staff I had licenses to use them at home. It is all too easy
to use such software, when there are site licenses for it, it is
installed on every machine.
5) You say "Every successful open source project has established itself
as THE alternative. Firefox is THE choice of browser after IE (who
really has heard of Opera?) OpenOffice is THE office suite etc. This
isn't true yet in the science software. Two must die to give the third
any recognition, and if MuPAD were to join the group, the problem would
get worse."
But I don't see that being any different to commercial software.
Word became popular as the alternative to word perfect, that was by far
the most popular word processor at one time.
Paint Shop Pro is commercial, but is a low-end alternative to Photoshop.
If the market for a product is large, then it will have a lot of
companies competing for a share of that market. If it is very obscure,
then it will not. That applies to anything - not just software.
--
Dave K
Minefield Consultant and Solitaire Expert (MCSE).
Please note my email address changes periodically to avoid spam.
It is always of the form: month-year@domain. Hitting reply will work
for a couple of months only. Later set it manually.
Richard Fateman
>
>
> To come back on topic, as along contributor to Macsyma you should be
> well positioned to understand some major drawbacks of commercial CAS.
> Macsyma is a prime example of wasted efforts having deserved the field,
> that would have been avoided with the GPL.
You presume to be able to predict the future as well as the hypothetical
past, and to be
able to compare the relative merits of GPL (perhaps non-existent at the
time
government funding stopped circa 1978), and BSD (also non-existent
at the time), the DOE technology transfer [open source but not transferable]
and MIT's technology transfer ideas, such as they were.
I think MIT totally botched the technology transfer to
Symbolics .
I worked hard to get a copy in the DOE library, but
my success was limited by the fact that the software, once taken
out of the library, could not be legally transferred to others.
One can ask speculate how Macsyma would have evolved
if it had been released under the same license as Berkeley BSD
along with Franz Lisp. Or under GPL.
One can wonder what Macsyma would look like now -- note that
Maple came along in about 1982 -- in some ways a re-engineering
of Macsyma ; Mathematica happened, too.
The internet, then arpanet, was not available to most peoples'
homes unless they were connecting to a universityi
My point, and the one you seem to be missing again,
is that you claim a hypothetical benefit from GPL.
But you neglect to speculate about whether it is better than BSD licensing,
or LGPL licensing, or ... A view that GPL=good,
non-GPL=bad is overly simplistic.
My guess is that for the technology transfer "experts"
at the MIT laboratory for computer science, and for
their "agents" at Arthur D. Little, they may still think
what they did was really great. They squeezed money out
of a project. Symbolics may actually have made money
on Macsyma, for some of the time. The people they employed
earned salaries, etc. If Macsyma has been made public,
I doubt that any people would have bought houses, raised
children, etc. with the fruits of their knowledge about
Macsyma. Would the mostly-pre-internet world have
just abandoned it? Dunno.
You could ask your questions about tech transfer, MIT and
GPL for Macsyma directly to Professor Joel Moses.
I doubt he will answer.
I've made my position pretty clear that I would have preferred
a BSD license at the time, but frankly, I cannot predict the
hypothetical past.
RJF
DP
...
> My point, and the one you seem to be missing again,
> is that you claim a hypothetical benefit from GPL.
> But you neglect to speculate about whether it is better than BSD licensing,
> or LGPL licensing, or ... A view that GPL=good,
> non-GPL=bad is overly simplistic.
Did I say that? No. It is not because I think that GPL would
(notice conditional) have prevented Macsyma disappearing that
I must assume that other licenses would not. But all the known
licenses that would have protected more or less effectively
Macsyma from financial hazards are certainly open source.
The point was not to make history-fiction but to learn
from the Macsyma episode how to license math software now.
I am glad that Axiom was finally released by NAG with the
BSD/GPL licenses.
Otherwise thanks for the interesting pieces of information
about Macsyma.
Dan
Richard J. Fateman
DP wrote:
> Richard Fateman wrote:
> ...
>
>> My point, and the one you seem to be missing again,
>> is that you claim a hypothetical benefit from GPL.
>> But you neglect to speculate about whether it is better than BSD
>> licensing,
>> or LGPL licensing, or ... A view that GPL=good,
>> non-GPL=bad is overly simplistic.
>
>
> Did I say that? No. It is not because I think that GPL would
> (notice conditional) have prevented Macsyma disappearing that
> I must assume that other licenses would not. But all the known
> licenses that would have protected more or less effectively
> Macsyma from financial hazards are certainly open source.
I think that perhaps you are not using GPL in its technical
sense. But I am not sure.
GPL is not the same as open source.
For example, from
http://opensource.org/docs/definition.php
"The license must not place restrictions on other software that is
distributed along with the licensed software. For example, the license
must not insist that all other programs distributed on the same medium
must be open-source software."
GPL places such restrictions.
>
> The point was not to make history-fiction but to learn
> from the Macsyma episode how to license math software now.
Given the changes in computing technology and networking,
it is not clear that the lessons from that time can be
applied today.
> I am glad that Axiom was finally released by NAG with the
> BSD/GPL licenses.
I don't know what you are referring to. Axiom was not released
by NAG under GPL. BSD is not GPL.
see http://directory.fsf.org/math/axiom.html
Jay Belanger
"Richard J. Fateman" <fat...@eecs.berkeley.edu> writes:
...
> GPL is not the same as open source.
>
> For example, from
> http://opensource.org/docs/definition.php
>
> "The license must not place restrictions on other software that is
> distributed along with the licensed software. For example, the license
> must not insist that all other programs distributed on the same medium
> must be open-source software."
>
> GPL places such restrictions.
No it doesn't; I have no idea where you'd get that idea.
The GPL itself says there are no such restrictions.
Jay
Richard J. Fateman
Jay Belanger wrote:
It does? Can you point to a specific place?
I wonder if we are reading the same material. I am not a lawyer, but
http://www.gnu.org/copyleft/gpl.html
condition 2.b. says
You must cause any work that you distribute or publish, that in whole or
in part contains or is derived from the Program or any part thereof, to
be licensed as a whole at no charge to all third parties under the terms
of this License.
and also, from the preamble,
For example, if you distribute copies of such a program, whether
gratis or for a fee, you must give the recipients all the rights that
you have. You must make sure that they, too, receive or can get the
source code. And you must show them these terms so they know their rights.
Note that there are lots of pieces of software, especially scientific
software, that are open-source, free, and not GPL. See for example
netlib.org
The idea that GPL is the only or best or even advisable way of
licensing code is, in my view, a misconception. But again, I am
not a lawyer. But I suspect that you are not a lawyer either.
Pondering this, all I can come up with is that you and I disagree about
the meaning of the term "restriction".
Say we are neighbors and I have a local wireless network. You use
it free.
1. Would it be a restriction on me if the law REQUIRED that I allow you
to use it?
2. Would it be a restriction on me if the law required that I allow you
to swim in my swimming pool?
3. Would it be a restriction on me to require that you be allowed make
free copies of programs that I wrote?
I say, yes, each of those are restrictions on my rights. Stallman
very deliberately wants to limit the rights of programmers to own their
own programs. He wants to restrict the rights of anyone
who uses GPL software to forbid them from redistribution unless they
wish to make their own code public. If you do not see that as
a restriction, then I think we have to agree to disagree.
RJF
RJF
DP
of the license, which is not completely unambiguous indeed, but which
was clarified in numerous occasions:
Richard J. Fateman wrote:
>
...
> condition 2.b. says
>
> You must cause any work that you distribute or publish, that in whole or
> in part contains or is derived from the Program or any part thereof, to
> be licensed as a whole at no charge to all third parties under the terms
> of this License.
The crucial word is "contains". What contains the Program or the derived
Program? Is it the whole netlib.org library, a whole DVD, an application,
or a file, if anyone of these contains some GPL code?
There was also discussions about dynamically linked libraries, should such
libraries be GPLed if the calling program is GPL, or vice versa?
Such issues should be clarified in GPL v 3.
Anyway in practice the container is an application, made of a group of files
being linked in a single binary file.
For example:
1) Most Linux distributions include 1000's of variously licensed programs and
files on the same DVD.
2) A computer vendor can sell a server running a GPL Linux kernel interacting
with proprietary or open source programs.
> Stallman very deliberately wants to limit the rights of programmers to own their
> own programs. He wants to restrict the rights of anyone
> who uses GPL software to forbid them from redistribution unless they
> wish to make their own code public. If you do not see that as
> a restriction, then I think we have to agree to disagree.
Congratulation, this piece misrepresents the GPL as well as
Microsoft propaganda.
In fact Stallman very deliberately wants to preserve the freedom of
program users to modify, run and distribute the GPL code
they have freely accepted as gift from the program authors.
The original authors have of course been free to select the
GPL license. Nobody is forced to run, to distribute or to
modify GPL programs.
This is the best compromise giving the maximum of freedom
to program authors as well as to users. The only restriction
is that anyone distributing GPL derived code should concede
the same privileges to others that the received ones.
Pretty civilized it seems. Yet RJF finds this a serious
restriction. This is indeed a major restriction for
companies wanting to reap profit of GPL code without return...
Dan
DP
Richard J. Fateman wrote:
...
>> I am glad that Axiom was finally released by NAG with the
>> BSD/GPL licenses.
>
> I don't know what you are referring to. Axiom was not released
> by NAG under GPL. BSD is not GPL.
>
> see http://directory.fsf.org/math/axiom.html
>
The version I have tells me the license is BSD/GPL, which means that
the axiom package distributed by Mandriva has both licenses.
---------------------------------------------------------------------------
$ rpm -qi axiom
Name : axiom Relocations: (not relocatable)
Version : 3.4 Vendor: Mandriva
Release : 0.20050901.1mdk Build Date: jeu 20 oct 2005 15:28:10
CEST
Install Date: ven 28 oct 2005 09:51:20 CEST Build Host: n5.mandriva.com
Group : Sciences/Mathematics Source RPM: axiom-3.4-0.20050901.1md
k.src.rpm
Size : 92505556 License: BSD/GPL
Signature : DSA/SHA1, sam 22 oct 2005 19:20:21 CEST, Key ID dd684d7a26752624
Packager : Giuseppe Ghibò <gh...@mandriva.com>
URL : http://axiom.axiom-developer.org
Summary : Symbolic Computation Program
Description :
Axiom is a general purpose Computer Algebra system.
It is useful for research and development of mathematical algorithms.
It defines a strongly typed, mathematically correct type hierarchy.
It has a programming language and a built-in compiler.
Richard Fateman
It is fairly clear that the original Axiom people did not want to
place such restrictions on other people. and so of course anyone
can ignore your version and get a copy from the site I
mentioned.
"DP" <pfen...@obs.unige.ch> wrote in message
news:440ce24d$1...@nntp.unige.ch...
>,,,, snip.....> The version I have tells me the license is BSD/GPL, which
Jay Belanger
"Richard Fateman" <fat...@cs.berkeley.edu> writes:
> Anyone can take a BSD program and re-release it under GPL.
> It is fairly clear that the original Axiom people did not want to
> place such restrictions on other people. and so of course anyone
> can ignore your version and get a copy from the site I
> mentioned.
I think that the complete Axiom distribution contains many pieces
under many different licenses. The NAG released part (the part that
makes it Axiom, I suppose) is BSD licensed, but now it contains parts
under different licenses.
> "DP" <pfen...@obs.unige.ch> wrote in message
> news:440ce24d$1...@nntp.unige.ch...
>>,,,, snip.....> The version I have tells me the license is BSD/GPL, which
>>means that
>> the axiom package distributed by Mandriva has both licenses.
Perhaps it means that parts are BSD licensed, parts are GPLed.
Jay
Richard Fateman
> The main point of RJF misunderstanding about the GPL comes from this part
> of the license, which is not completely unambiguous indeed, but which
> was clarified in numerous occasions:
>
> Richard J. Fateman wrote:
>>
> ...
>> condition 2.b. says
>>
>> You must cause any work that you distribute or publish, that in whole or
>> in part contains or is derived from the Program or any part thereof, to
>> be licensed as a whole at no charge to all third parties under the terms
>> of this License.
>
> The crucial word is "contains". What contains the Program or the derived
> Program? Is it the whole netlib.org library, a whole DVD, an application,
> or a file, if anyone of these contains some GPL code?
Does Maxima contain Lisp? Does Lisp contain Maxima? Does Maxima
contain a front-end graphics program? Or an intertwined (call-back) plotting
program or numerical library? I am not talking about sending someone
a DVD with incidentally a copy of emacs on it.
> There was also discussions about dynamically linked libraries, should such
> libraries be GPLed if the calling program is GPL, or vice versa?
> Such issues should be clarified in GPL v 3.
You mean the one that Linus Torvalds refuses to use?
http://www.theregister.co.uk/2006/01/26/linux_torvalds_gpl/
>
> Anyway in practice the container is an application, made of a group of
> files
> being linked in a single binary file.
> For example:
> 1) Most Linux distributions include 1000's of variously licensed programs
> and
> files on the same DVD.
The view of the world as a collection of programs called from the UNIX
shell
was pretty good in 1975. It doesn't cover the situation now.
> 2) A computer vendor can sell a server running a GPL Linux kernel
> interacting
> with proprietary or open source programs.
I believe explicit exceptions are made for operating systems, so that you
can
run a GPL program under Windows.
Lisp is not usually considered an operating system.
>
>> Stallman very deliberately wants to limit the rights of programmers to
>> own their
>> own programs. He wants to restrict the rights of anyone
>> who uses GPL software to forbid them from redistribution unless they
>> wish to make their own code public. If you do not see that as
>> a restriction, then I think we have to agree to disagree.
>
> Congratulation, this piece misrepresents the GPL as well as
> Microsoft propaganda.
Congratulations for using ad hominem argumentation.
>
> In fact Stallman very deliberately wants to preserve the freedom of
> program users to modify, run and distribute the GPL code
> they have freely accepted as gift from the program authors.
--- some of whom may not have wanted to offer their programs, but
--- were coerced, by the GPL juggernaut. They did not
--- want to have to rewrite some piece of code they used. This
--- is precisely Stallman's strategy. It is effective; Stallman
--- is no fool.
> The original authors have of course been free to select the
> GPL license.
Nobody is forced to run, to distribute or to
> modify GPL programs.
No, you are perfectly free to start from an empty directory
and recreate a computer algebra system.
> This is the best compromise giving the maximum of freedom
> to program authors as well as to users.
What other compromises have you explored?
The maximum freedom to program authors would allow them
to write programs and then (a) give them away; (b) sell them;
(c) provide source; (d) not provide source. (e) license them.
or anything else they wish. Now if the programs use other
peoples programs in some way, those other programs could
be bought, licensed (perhaps free), adopted from public domain,
etc.
> The only restriction
> is that anyone distributing GPL derived code should concede
> the same privileges to others that the received ones.
Not really. Your rights in the program you write are poisoned
by using/including in your executable program
a GPL program X. This is Stallman's intention. You are
of course free to write another (proprietary) version of X that
you use, and then you can own it. I think that it is
far more civilized in the first place, since I am paid as
an academic, to hand someone my program to use as they
see fit. If I had to live by my programming, I would find
it extremely distasteful to be restricted by GPL. Yes of course
it is possible to sell free programs to a willing buyer. Once?
> Pretty civilized it seems. Yet RJF finds this a serious
> restriction. This is indeed a major restriction for
> companies wanting to reap profit of GPL code without return...
It is a far more significant restriction on USERS who would
like to see an environment in which GPL and other programs
can coexist and interact in a modern workstation.
Today's workstations usually are not restricted to
"shells" and "pipes" working at arms' length.
There are other models for rights in programs, and GPL is rather
complicated and restrictive, I think. LGPL is better, in my view;
BSD is better also. If GPL were so great, why is it so so
complicated? Apparently open to controversy. In its 3rd version.
(so far as I can tell) untested by courts.
And this is probably still mostly off topic for sci.math.symbolic.
I can at least email Jay offline.
RJF
Richard Fateman
>
>
> Perhaps it means that parts are BSD licensed, parts are GPLed.
>
In that case I would suggest that anyone using Axiom ignore any copy
under GPL, and get the original unrestricted Axiom code.
The authors had a different intention, and it seems to me
kind of rude to try to restrict the distribution of the code by telling
people that it is GPL.
RJF
DP
Richard Fateman wrote:
...
> Does Maxima contain Lisp? Does Lisp contain Maxima?
> Does Maxima contain a front-end graphics program?
> Or an intertwined (call-back) plotting
> program or numerical library?
The whole package is declared GPL, no problem.
___________________________________________________________________________________
$ rpm -qi maxima
Name : maxima Relocations: (not relocatable)
Version : 5.9.2 Vendor: Mandriva
Release : 1mdk Build Date: lun 07 nov 2005 13:39:47 CET
Install Date: mar 08 nov 2005 12:18:17 CET Build Host: n4.mandriva.com
Group : Sciences/Mathematics Source RPM: maxima-5.9.2-1mdk.src.rpm
Size : 17787654 License: GPL
Signature : DSA/SHA1, lun 07 nov 2005 15:20:22 CET, Key ID dd684d7a26752624
Packager : Giuseppe Ghibò <gh...@mandriva.com>
URL : http://maxima.sourceforge.net
Summary : Maxima Symbolic Computation Program
Description :
Maxima is a full symbolic computation program. It is full featured
doing symbolic manipulation of polynomials, matrices, rational
functions, integration, Todd-coxeter, graphing, bigfloats. It has a
symbolic debugger source level debugger for maxima code. Maxima is
based on the original Macsyma developed at MIT in the 1970's. It is
quite reliable, and has good garbage collection, and no memory leaks.
It comes with hundreds of self tests.
-------------------------------------------------------------------------------------
...
>>> Stallman very deliberately wants to limit the rights of programmers to
>>> own their
>>> own programs. He wants to restrict the rights of anyone
>>> who uses GPL software to forbid them from redistribution unless they
>>> wish to make their own code public. If you do not see that as
>>> a restriction, then I think we have to agree to disagree.
>> Congratulation, this piece misrepresents the GPL as well as
>> Microsoft propaganda.
> Congratulations for using ad hominem argumentation.
Wrong, I criticize your misrepresenting text, no unrelated characteristics
of your person.
Wikipedia: "An ad hominem fallacy consists of asserting that someone's
argument is wrong and/or they are wrong to argue at all purely because of
something discreditable/not-authoritative about the person or those persons
cited by them rather than addressing the soundness of the argument itself."
(...)
> And this is probably still mostly off topic for sci.math.symbolic.
Yes, I will be glad to stop replying to your postings.
Dan
Jay Belanger
"Richard Fateman" <fat...@cs.berkeley.edu> writes:
...
>> Perhaps it means that parts are BSD licensed, parts are GPLed.
>>
>
> In that case I would suggest that anyone using Axiom ignore any copy
> under GPL, and get the original unrestricted Axiom code.
I would advise differently. Tim Daly has put a lot of work into
getting the freed Axiom working nicely; I intend to use and respect
what he's done.
> The authors had a different intention, and it seems to me kind of
> rude to try to restrict the distribution of the code by telling
> people that it is GPL.
It would also be rude to ignore the intentions of someone who put
their code under the GPL, as you've been doing to Bill Schelter's
work. Try to show the same decency that you claim to expect from
others.
Jay
Richard J. Fateman
Jay Belanger wrote:
>
> It would also be rude to ignore the intentions of someone who put
> their code under the GPL, as you've been doing to Bill Schelter's
> work. Try to show the same decency that you claim to expect from
> others.
>
A glance at one maxima 5.9.0 source directory shows some 100,000 lines
of lisp code by many authors. Most of that code can be obtained,
though somewhat inconveniently, from DOE under what amounts to a BSD
license.
The DOE license explicitly allows adoption of the code in any way,
including incorporation into products. The DOE gave Bill non-exclusive
rights to distribute under GPL; others presumably can request the
right to distribute in other ways.
I agree that one should generally not ignore the intentions of the
Maxima authors. Sometimes these intentions are not clear, and in any
case the MIT authors turned over rights to DOE, so DOE's intentions
matter too.
If Bill Schelter's own programs (indeed some very nice ones!) are
treated as GPL because he explicitly said so, that's fine. Though if
he gave them to the DOE (for some code, he did...), then DOE's
intentions matter too. Stamping GPL on them AFTER giving to DOE... hmm.
For the code that he did not write, we can either (a) look to the
authors or (b) look to the DOE which apparently has rights -- the DOE
paid for the research-- or (c) disregard the authors and do whatever
we decide to do.
As one of those authors, it is NOT my intention to restrict these
programs by the covenants in the GPL.
So if you really want to respect the authors intentions, you might
advocate the clear separation of the GPL-at-birth parts of Maxima from
the GPL-by-association parts (most of it). Any executable code package
with BOTH kinds of components would necessarily be GPL. You can still
run something like Maxima without any GPL parts.
A package without GPL-at-birth parts would be anything
consistent with DOE's license.
RJF
Jay Belanger
> Maxima authors.
Unless it doesn't suit your purposes, apparently.
> So if you really want to respect the authors intentions, you might
> advocate the clear separation of the GPL-at-birth parts of Maxima from
> the GPL-by-association parts (most of it).
Now you're playing games.
And even for the parts that Bill Schelter wrote after he got the code
from the DOE, you have suggested ignoring his intentions.
If you want a copy of Maxima that isn't GPLed, you need to go to DOE
to get it. Unless respecting the intentions of Bill Schelter doesn't
matter, of course.
> A package without GPL-at-birth parts would be anything
> consistent with DOE's license.
Huh?
At any rate, a GPLed Maxima is what DOE gave permission for.
Jay
Dave (from the UK)
> If GPL were so great, why is it so so
> complicated?
I would agree the GPL is a bit long-winded, but you like myself are a
scientist, so are probably not used to dealing with legal documents. I
suspect it is not untypical for such a document. It was drafted by
lawyers in the US to stand up in court. It seems to have done so in
Europe at least, so I think they did a reasonable job.
> Apparently open to controversy.
It %$£$£! Microsoft off, which can only be a good thing!!
> In its 3rd version.
Given the age of it, that is not unreasonable. The computer industry
moves quickly. You are not obliged to use the latest revision.
> (so far as I can tell) untested by courts.
The GPL has been tested in court in Europe and did its job.
Harald Welte, one of the main netfilter authors, sued a Dutch company,
Sitecom, alleging it used the software in a wireless network product
without abiding by the terms of the General Public License (GPL).
http://news.com.com/GPL+gains+clout+in+German+legal+case/2100-7344_3-5198117.html
http://www.theregister.co.uk/2004/04/21/licence_germany/
That was later upheld in a final hearing on the matter.
If you do a Google on:
Sitecom GPL
you will get many hits on this. Not surprisingly, if you go to the
Sitecom web site
and type in GPL, you will not get a single hit.
I am sure it is abused in many cases, and the problem is the authors of
the code often can't afford a legal bill to take on a company.
It some cases it would be hard to know too. Someone contacted me, asking
if I would change the license on
from GPL to LPL so they could include it in their commercial PCB
software. I declined, but whilst I have no reason to believe they are
using it, quite honestly if they did, I would most unlikely to be able
to prove it. And if I did know, there is nothing much I could do about
it, other than slag them off on a web page (and that would be
financially risky for me).
Waldek Hebisch
>
>
> Richard J. Fateman wrote:
> ...
> >> I am glad that Axiom was finally released by NAG with the
> >> BSD/GPL licenses.
> >
> > I don't know what you are referring to. Axiom was not released
> > by NAG under GPL. BSD is not GPL.
> >
> > see http://directory.fsf.org/math/axiom.html
> >
>
> The version I have tells me the license is BSD/GPL, which means that
> the axiom package distributed by Mandriva has both licenses.
>
Well, explanation is simple: Axiom is build on top of GCL. Part of GCL
(the readline and BFD libraries) are GPL (rest of GCL is LGPL). So
axiom binaries fall under GPL too. Pure Axiom sources are BSD, but
last time I checked distribution tarball bundled GCL, so technically
it was both under BSD and GPL.
--
Waldek Hebisch
heb...@math.uni.wroc.pl
Waldek Hebisch
>
> GPL is not the same as open source.
>
> For example, from
> http://opensource.org/docs/definition.php
>
> "The license must not place restrictions on other software that is
> distributed along with the licensed software. For example, the license
> must not insist that all other programs distributed on the same medium
> must be open-source software."
>
> GPL places such restrictions.
>
GPL end of Section 2:
In addition, mere aggregation of another work not based on the
Program with the Program (or with a work based on the Program) on
a volume of a storage or distribution medium does not bring the
other work under the scope of this License.
So, please check your claims with facts. You may have valid points
agaist GPL, but making false claims only weakens your arguments.
--
Waldek Hebisch
heb...@math.uni.wroc.pl