On Tue, Apr 21, 2009 at 10:34 PM, Kustaa Nyholm
<[email protected]> wrote:
>>> I hope I did not talk about any compiler, that was not my intention. I was
>>> just saying that this GPLv2 clearly states that a compiler can be a major
>>> component of a system.
>>>
>>
>>Yes, it can - for instance the unix and cc. But I don't think this
>>applies to MSVC++ and Windows.
> I'm not agreeing or disagreeing but what is the reasoning why Windows / MSCV
> is different from say Linux/gcc?
>
>>If you consider the library not part of MSVC++, then it must be a
>>separate "component", therefore explicitly disallowed to be
>>distributed along ("...unless that component itself accompanies the
>>executable").
>
> It does not follow that some thing that is not a component is itself a
> component.
> The license stipulates that there can be something that "is normally
> distributed
> (in either source or binary form) with the major components".
If the "anything" word refers to MSVC++ runtime library, then it must
be part of a component. And you said it's not part of MSVC++. So what
component is it part of? To me, the only logical consequence is that
you consider the runtime to be separate.
> And this paragraph really tries to specify what that something and that this
> something
> is exempt from the source code distribution. So it does not follow that it is
> explicitly disallowed, on the contrary.
>
> The contention is weather MSVC++ is a component of the
> operating system and if the libraries are "something that is normally
> distributed with the major components" and weather the libraries
> are part of MSVC++ or not.
>
>>I meant to say that I don't think you can cheat the GPL around by me
> re
>>breaking the "component" into parts (and possibly omitting some).
> I had no intention to cheat, on the contrary, my intention is, if anything,
> to support Free software with an interpretation of GPL that seem to
> be in line with the stated goals of GPL creators. Not arguing to
> make it more difficult for people to use Free software in real world
> where we need to mix and mingle with non Free software.
>
Sorry, but that's what the GPL is mostly about - to make mixing with
proprietary software more difficult.
GPL software just favors other GPL (and free) software. To guote RMS,
"Proprietary software developers have the advantage of money; free
software developers must provide advantage fro each other".
>
>>>And in my view, the intent was to allow linking to them, not
>>>distributing them along (hence the last sentence).
>
> Sorry now this confused me totally. I thought I was the one that
> was advocating for allowing linking them and you were for distributing
> them along unlinked?
>
Then there was indeed a confusion. I think what the installer does is
that it distributes the octave binaries that are linked to MSVC++
*dynamic* runtime libs (DLLs) and it also distributes these DLLs. And
that's the problem.
If the MSVC++ runtime library is a component of Windows, then it
should not be distributed ("accompany the executable"), but the Octave
binary can. If not, then the Octave binary can't.
Now, if you substitute in an *arbitrary* proprietary library, then
that's OK - this is what the GPL is supposed to do.
Of course, the intent even in GPL2 was probably to exclude all
compiler runtimes, but GPL2 fails to do so due to its wording.
> Distributing them along would seem to be something that the
> GPL allows: "mere aggregation of another work <snip> does not bring the
> other work under the scope of this License."
Yes, but here it's linking *and* aggregation, and that seems not allowed.
> Or maybe I was (intentionally ;-) miss interpreting your words?
>
>>Well, GPLv2 does not define what "component" is in any way, so the
>>implication "we cannot regard... because the license does not say it"
>>is wrong.
>
> Well, it gives two examples:kernerl and compiler so at least something
> is defined of what a component of the operating system is.
>
Examples are not definitions. I think it is apparent that these were examples.
GCC can be considered a component of Debian because it can be selected
in installation, or downloaded via package manager. But you can't find
MSVC++ on the windows DVD nor you can get it through window update.
And there is otherwise no sign that MSVC++ would be a component of
windows, other than the fact that they're both from Microsoft. I ask
again: Is Salford Fortran component of Windows? If not, then what's
the difference?
>
>>Yes, but I don't think MSVC++ is a component of an operating system.
>>Back to the start.
>
> Indeed, it depends on that. We have both stated our opposing views
> and argument why we think it is / it isn't. I, of course, think that
> I had some arguments that had at least some support in the text
> of the license, whereas I found your arguments less compelling
> and based less on the language of the agreement. You, no doubt, would
> beg to differ.
>
The only support you have shown is an *example* (in the parentheses)
that I don't think applies to this situation.
And I think that if you take two separate programs, Windows and
MSVC++, then by all common sense you must assume neither is component
of the other *unless there is evidence to the contrary*. Which you
haven't shown.
>> I'm saying that GPL license takes the view that the compiler that is used to
>> build
>> the program being licensed is a component of the operating system under which
>> the program is to be executed.
>
>>And I disagree. It doesn't say that.
>
> It doesn't say what? It clearly says that compiler is a component of the
> operating system,
> are you disputing that?
Yes. And without mathematical quantifiers, that sentence you gave does
not even make precise sense. So, pick yourself (all possibilities):
"Every compiler is a component of any operating system (on which it runs)"
"At least one compiler is a component of any operating system it runs on"
"Every compiler is a component of at least one operating system it runs on"
"At least one compiler is a component of at least one operating system
it runs on"
so, which one is, according to you, stated by GPL2?
> And if that is any compiler why would it be mentioned at all in that clause?
> So to me it is clear that the compiler is the compiler that is involved in
> the making of the
> program that is being licensed under GPL.
I agree that was likely the intention, but "it's clear" doesn't help -
the wording says otherwise. The fact that "a compiler X is a component
of operating system Y" can't depend on whether you made any GPL
software using it. It is an independent statement, and for any pair
(X,Y) must be either true or false. So, according to you, X = MSVC++
and Y = Windows makes it true. Fine. Is X = Salford Fortran and Y =
MSVC++ also true?
GPL3 cleared up this problem. It clearly defines things.
--
RNDr. Jaroslav Hajek
computing expert & GNU Octave developer
Aeronautical Research and Test Institute (VZLU)
Prague, Czech Republic
url: www.highegg.matfyz.cz
------------------------------------------------------------------------------
Stay on top of everything new and different, both inside and
around Java (TM) technology - register by April 22, and save
$200 on the JavaOne (SM) conference, June 2-5, 2009, San Francisco.
300 plus technical and hands-on sessions. Register today.
Use priority code J9JMT32. http://p.sf.net/sfu/p
_______________________________________________
Octave-dev mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/octave-dev
