Re: [OctDev] [gnu.org #432927] Can a Windows installer include both VC++ libs and GPLed libs?

[Jaroslav Hajek via RT]

On Thu, May 21, 2009 at 10:49 AM, Kustaa Nyholm via RT
<address@hidden> wrote:
>
> Hi Brett,
>
> thanks for the answer. I'm not really a party to this issue but very
> interested anyway.
>
> Brett Smith wrote:
>
>> However, both versions of the GNU GPL have language which prevent this
>> exemption from applying to libraries that are actually distributed in
>> tandem with the GPLed software.
> Could you pin point the "language" here so that we (GPL users) would
> understand better what it is that this is based on.
>
>
>
>>
>> The reason the GPL works this way is because we need to keep the System
>> Library exception very narrow.  If we make it too easy for libraries to
>> qualify as System Libraries, it will become feasible for companies to
>> change free software and keep the changes proprietary by putting them in
>> a "System Library."
> Seems like a bad way of trying to achieve this goal. Would it not have
> been better to specify that the Free software cannot be changed so
> that it can only be used with proprietary libraries for which no
> Free alternatives are available?

Such a formulation is too vague to be of any use. Try making it more
explicit and you'll see it's not really that easy.

> Like in this instance, the GPL seems to work against it's stated goals.

I don't think so. I actually disagree with Brett in the sense that I
think GPL 3 (not GPL 2) allows the distribution in question. But
otherwise, I agree that System Library exception is to be kept narrow,
for the precise reasons he stated.

>> To the best of my knowledge, this exact situation is the only one that
>> presents problems for distributors.
> And problems for users, very much for the users!
>
>> As far as I'm aware, on every other
>> major operating system in use today, all the libraries that would
>> qualify as System Libraries come with the operating system, or are at
>> least part of the standard install.  The runtime libraries for
>> Microsoft's compilers are the only exception.
> It maybe the 'only' exception but  Windows is 95% or something of all 
> computer users, thinking otherwise is leaving in fantasy world, no matter how 
> much I dislike many of the business practices and things that M$ stands for.
>
> Also your sentence above seems to imply that actually runtime libraries
> for Microsoft's compilers would be System Libraries if they were
> distributed with the OS or were part of the standard install?
>
> Not a very coherent argument, especially as what we are talking about
> here is not some malicious attempt to hijack Free software for
> proprietary causes but for making life for Free software users
> much easier and thus the use of Free software more wide spread.
>
>
>>>
>> No.  Distributing the Octave binaries and the runtime libraries on the
>> same media would be problematic, regardless of the specific medium used.
> That is a very problematic sentence. What if the medium is a the net?
> Distributing on the same server is forbidden? On the same (Inter)net?
>
>>
>> I think the Windows binary distribution should simply provide users with
>> instructions to obtain the libraries from Microsoft's site.  I realize
>> that's inconvenient, but hopefully it's not too bad, and I think it's a
>> worthwhile change to avoid any GPL trouble.
> Yeah, it *is* too bad. Very, very few Windows users (and remember
> that is 95% of the potential user base) are able to do this and even if
> they are, they are not willing to do this. Too bad for GPLed software,
> good for proprietary software. Personally, being a mainly Mac OS X user,
> I have even a lower threshold. If I cannot double click on a disk image
> on webpage and run the software *without installing* it at all, there needs
> to be a very, very, very good reason for me to even try any software.
>

GPLed software needs to protect itself against proprietarization
(that's the point of GPL) and some sacrifices seem necessary to
achieve this goal. If you have a better idea for GPL4, preserving its
main goals, share it with the FSF.

>
>>
>> If you have further questions, please feel free to contact me; I'll be
>> on the lookout for those, and try to respond as quickly as possible.
> I think we all appreciate your efforts.
>
> Like I declared I have no real vested interest in octave (except a happy
> user on Mac OS X) but out of curiosity I talked to a lawyer about this
> and so for the record her opinion was that FSF is not really party the
> license agreement between those who release software under GPL and
> those who use the software. So it is pretty much up those to put
> the spin on the license in court if it comes to that, not FSF.
>
> As an ever increasing number of people contribute to a GPL licensed piece of 
> software, releasing their creative work under *their* assumption of what the 
> GPL means on any given date it is a horrible mess! Think of some
> developer testifying that 'I released my contribution under the assumption
> that <some assumption>... and now I'm detracting that contribution
> because it seems that GPL means <something else>, and I demand that
> all those users that are using the software stop using it'.
>

Every developer using GPL is supposed to read it and understand it. If
he relies instead on what he's told from other sources, he does so at
his own risk. I agree with your point that the FSF position to GPL is
not normative, only advisory. GPL is bound by its own wording; and
only courts are entitled to resolve disputes that can't be resolved
otherwise.

regards

-- 
RNDr. Jaroslav Hajek
computing expert & GNU Octave developer
Aeronautical Research and Test Institute (VZLU)
Prague, Czech Republic
url: www.highegg.matfyz.cz