FAQ Search Today's Posts Mark Forums Read
» Video Reviews

» Linux Archive

Linux-archive is a website aiming to archive linux email lists and to make them easily accessible for linux users/developers.


» Sponsor

» Partners

» Sponsor

Go Back   Linux Archive > Redhat > Fedora Development

 
 
LinkBack Thread Tools
 
Old 06-18-2008, 11:26 PM
"max bianco"
 
Default Fedora Freedom and linux-libre

On Wed, Jun 18, 2008 at 6:20 PM, Alexandre Oliva <aoliva@redhat.com> wrote:
> On Jun 18, 2008, max <maximilianbianco@gmail.com> wrote:
>
>>> Yes. Which is why we should stop supporting these "PCs" which aren't
>>> completely free.
>
>> I'm all for it too but first you have to get completely free options
>> on the market place.
>
> Why does it have to be 'first'? Why can't we build up pressure such
> that vendors will have more of a reason to do it?
>

Vendors are motivated by money, nothing more nothing less if you show
them they can save money *and* not lose clients they will go for it.
Inertia drives this market, the train is coming to a halt but it takes
time to stop it and then of course you have to change direction. There
is also the fact that most people don't know enough about the subject
to care so if you don't put it front of them and maintain the price
point they are used too then it is going to be especially difficult.
These things take time, things are starting to change, I'd like
everyone to suddenly wake up to what's going on around them too.


> --
> Alexandre Oliva http://www.lsd.ic.unicamp.br/~oliva/
> Free Software Evangelist oliva@{lsd.ic.unicamp.br, gnu.org}
> FSFLA Board Member íSÚ Libre! => http://www.fsfla.org/
> Red Hat Compiler Engineer aoliva@{redhat.com, gcc.gnu.org}
>



--
If opinions were really like assholes we'd each have just one

--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
 
Old 06-18-2008, 11:43 PM
Les Mikesell
 
Default Fedora Freedom and linux-libre

Alexandre Oliva wrote:


Here, take this program here:

http://www.lsd.ic.unicamp.br/~oliva/papers/vta/example.c

What do you think you're entitled to do with it as it stands?


Assuming it is permitted for me to get a copy in the first place, then I
have fair use of it. I could link my copy with anything else I wanted.
And I could give away my work that called it and some other code.



Now, if I tell you, I license it to you under the GPL, and you accept
it. what you you think you're no longer entitled to do with it?


Then I couldn't give my additional work away to other people to use with
their own copies of your work and third party libraries. They could
hire me to re-invent it, but I couldn't distribute my work.



No one but the FSF claims that a patch is a derived work


How would you defend the claim that something that literally copies
portions of a copyrightable work is not?


It's fair use, but I'd settle for being able to do a line-numbered edit
script that didn't not copy any original material. Or a completely
separate work that linked two differently licensed libraries that are
not included.



or a separate component that links 2 others together but doesn't
contain a copy of the others.


Even dynamic linking copies portions of the dynamic libraries. We've
covered that already.


Yes, we covered the fact that everyone else permits it.


Try to create a derived work based on say Microsoft Windows or
Microsoft Word, if you happen to have them around, and to distribute
it, and see what happens.



I'm not sure what you are talking about.


Try creating a work based on internal DLLs, for which you have no
"developer" permissions.


Again, I have never heard of any prohibition against sharing such a
thing with anyone who has their own copy of all required components.



It doesn't. Your own original work can't possibly be a derived work.



Difference of opinion, I guess. The FSF says otherwise


No, it doesn't, and the GPL itself confirms that. There's a
difference between "your own original work" and "the derived work you
created out of a pre-existing third-party's work". The above is about
the former, as you can probably tell from the similarity in spelling ;-)


if the resulting 'work as a whole' would be be a derived work,


Then it's not "your own original work"


Define whose work it is, then. Take the case of an original program
that links with two other libraries. It is entirely conceivable and
fairly likly that compatible GPL and non-GPL libraries will be available
and it is the end users choice at compile/run time. Who owns this work?


It is a matter of circumstance whether the alternatives exist or not and
you may not even know. And such a bizarre definition changes from day
to day. One day the RIPEM code was a derived work because it linked to
the gmp library. With no changes to the code itself, it became not a
derived work because the unrestricted fgmp library existed - even though
no one really used it.





Without the GPL, I can give you a copy of my original work that
links to that library


Only if the copyright holder of the library granted you permission to
distribute your derived work.


That's just not true.

--
Les Mikesell
lesmikesell@gmail.com

--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
 
Old 06-19-2008, 09:34 AM
Andrew Haley
 
Default Fedora Freedom and linux-libre

Les Mikesell wrote:
> Andrew Haley wrote:
>>
>>>> I've explained that the GPL prevents me from sharing original work
>>>> that links to both GPL and non-GPL libraries.
>>> And I've explained that it doesn't, and asked you to cite the passage
>>> of the GPL that prevents you from doing it. You haven't bothered to
>>> do it, and instead decided to keep insisting in this nonsensical
>>> claim.
>>
>> Actually, he did reply to that, but perhaps you didn't see it.
>>
>> He had an unfree program (the wattcp library) and GNU tar and
>> discovered that he wasn't allowed to ship a program that linked
>> these two together. He then said, gloriously, "please don't try
>> to say the problem was cause by those other licenses -
>> they did not prevent anyone else from getting copies..." :-)
>>
>> OK Les, I promise not to say it. However...
>
> I'm not sure what you are trying to imply here. I could redistribute
> copies of the other two components - but that's irrelevant since anyone
> could get them in source anyway (they were only unfree in the reverse
> way that the GPL redefines free to mean restricted).

Err, one library according to you, was unfree in the sense that you weren't
allowed to change it in any way; to enhance it, or to fix bugs.

> I could have
> redistributed the combination if I had started with the original pdtar
> instead of gnutar.

You could, but if what you said was true, you still wouldn't be able to
fix bugs in one of the libraries. And if, to you, free software is that
which is free (as in beer) but you aren't allowed to fix bugs, then
you're welcome to it.

Andrew.

--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
 
Old 06-19-2008, 01:50 PM
Les Mikesell
 
Default Fedora Freedom and linux-libre

Andrew Haley wrote:

>

He had an unfree program (the wattcp library) and GNU tar and
discovered that he wasn't allowed to ship a program that linked
these two together. He then said, gloriously, "please don't try
to say the problem was cause by those other licenses -
they did not prevent anyone else from getting copies..." :-)

OK Les, I promise not to say it. However...

I'm not sure what you are trying to imply here. I could redistribute
copies of the other two components - but that's irrelevant since anyone
could get them in source anyway (they were only unfree in the reverse
way that the GPL redefines free to mean restricted).


Err, one library according to you, was unfree in the sense that you weren't
allowed to change it in any way; to enhance it, or to fix bugs.


That's not an accurate description, although it did have a restriction
on distributing modified versions. I could, of course, change my own
copy and submit bug fixes and enhancements to the author for
incorporation - or make the source modifications available separately
from the package. The restriction was more about preventing broken
versions from being distributed than enhancements.



I could have
redistributed the combination if I had started with the original pdtar
instead of gnutar.


You could, but if what you said was true, you still wouldn't be able to
fix bugs in one of the libraries. And if, to you, free software is that
which is free (as in beer) but you aren't allowed to fix bugs, then
you're welcome to it.


The best way to fix bugs is to get the fix incorporated in the base
package so it can be properly tested, merged with other fixes, and
provided to everyone using it. Otherwise you can s/fix/add/ in your
description of what you want to be permitted to do followed by
distributing them in a package someone else tries to support. How many
broken TCP stacks do we really want on our networks where someone thinks
cranking the retry timer gives it an advantage? There is the argument
that if the author/maintainer stops updating, the package can die.
Well, these were DOS libraries. The sooner that stuff died the better.
Knowledge about TCP didn't start or end with that package and copies
are probably still available in archives somewhere.


--
Les Mikesell
lesmikesell@gmail.com

--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
 
Old 06-19-2008, 01:57 PM
Andrew Haley
 
Default Fedora Freedom and linux-libre

Les Mikesell wrote:
> Andrew Haley wrote:
>> >
>>>> He had an unfree program (the wattcp library) and GNU tar and
>>>> discovered that he wasn't allowed to ship a program that linked
>>>> these two together. He then said, gloriously, "please don't try
>>>> to say the problem was cause by those other licenses -
>>>> they did not prevent anyone else from getting copies..." :-)
>>>>
>>>> OK Les, I promise not to say it. However...
>>> I'm not sure what you are trying to imply here. I could redistribute
>>> copies of the other two components - but that's irrelevant since anyone
>>> could get them in source anyway (they were only unfree in the reverse
>>> way that the GPL redefines free to mean restricted).
>>
>> Err, one library according to you, was unfree in the sense that you
>> weren't
>> allowed to change it in any way; to enhance it, or to fix bugs.
>
> That's not an accurate description, although it did have a restriction
> on distributing modified versions. I could, of course, change my own
> copy and submit bug fixes and enhancements to the author for
> incorporation - or make the source modifications available separately
> from the package. The restriction was more about preventing broken
> versions from being distributed than enhancements.

Sure, but it had that effect, didn't it? If you're not allowed to
distribute modified versions without someone else's consent, it's not free
(as in freedom) software.

>>> I could have
>>> redistributed the combination if I had started with the original pdtar
>>> instead of gnutar.
>>
>> You could, but if what you said was true, you still wouldn't be able to
>> fix bugs in one of the libraries. And if, to you, free software is that
>> which is free (as in beer) but you aren't allowed to fix bugs, then
>> you're welcome to it.
>
> The best way to fix bugs is to get the fix incorporated in the base
> package so it can be properly tested, merged with other fixes, and
> provided to everyone using it. Otherwise you can s/fix/add/ in your
> description of what you want to be permitted to do followed by
> distributing them in a package someone else tries to support. How many
> broken TCP stacks do we really want on our networks where someone thinks
> cranking the retry timer gives it an advantage? There is the argument
> that if the author/maintainer stops updating, the package can die.

Quite. And, indeed, that's the inevitable consequence.

Andrew.

--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
 
Old 06-19-2008, 03:36 PM
Les Mikesell
 
Default Fedora Freedom and linux-libre

Andrew Haley wrote:



Err, one library according to you, was unfree in the sense that you
weren't
allowed to change it in any way; to enhance it, or to fix bugs.

That's not an accurate description, although it did have a restriction
on distributing modified versions. I could, of course, change my own
copy and submit bug fixes and enhancements to the author for
incorporation - or make the source modifications available separately
from the package. The restriction was more about preventing broken
versions from being distributed than enhancements.


Sure, but it had that effect, didn't it? If you're not allowed to
distribute modified versions without someone else's consent, it's not free
(as in freedom) software.


It's not free the way the GPL redefines the word to mean restricted, but
it doesn't interfere with your freedom to distribute your changes as
patches, leaving it clear that it is something different from the
original author's work that he supports. In more modern licenses, I'd
prefer the ones where you are permitted to modify independently and
distribute the forked copy if you change the package name, but it is
only in odd circumstances that it even matters or that there is any
effective difference. Even in GPL circles I think most people agree
that the best process is to coordinate modifications into a single
revision tree instead of forking wildly.



There is the argument
that if the author/maintainer stops updating, the package can die.


Quite. And, indeed, that's the inevitable consequence.


It's not at all inevitable since the copyright holder can transfer
control at any time or might already be a foundation that will outlast
any possible use for the product. But, in technology everyone is better
off when an old package does die and is replaced by something new and
improved, and the harm of the GPL is that it's 'work as a whole'
requirement makes it difficult or impossible for these replacements to
happen at the component level when the currently best component isn't
encumbered by the GPL.


--
Les Mikesell
lesmikesell@gmail.com

--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
 
Old 06-19-2008, 03:59 PM
Alexandre Oliva
 
Default Fedora Freedom and linux-libre

On Jun 18, 2008, Matthew Saltzman <mjs@clemson.edu> wrote:

> Then can we at least agree that there are sometimes unfortunate
> consequences to the GPL's failure to permit one to share a work
> combining two pieces of *free* software because of relatively minor[1]
> license incompatibilities?

Yeah, it's unfortunate when this happens. In general, authors who use
the GPL for its intended purpose (ensuring the 4 freedoms are
respected for all users) won't object to the combination of their
works with other works that respect users' freedoms, and will grant
additional permissions for the combinations in spite of the license
conflicts.

Of course, not everyone does that, and some people who would like to
create such combinations may even not realize that this possibility
exists, or think it's not worth the effort.

So, yeah, it's unfortunate, but I don't think it's really such a big
deal. Nearly all Free Software *is* available under the GPL and
compatible licenses anyway.

> In fact, I think it's arguable that there are sometimes unfortunate
> consequences to the GPL's failure to permit one to share a work that
> makes use of a GPL library and a proprietary library.

Sparing a user from becoming dependent on a piece of proprietary
software might even be a sacrifice for the user, but it's actually an
advantage for the user and for society in the long run.

Anyhow, it is true that conquering freedom is not without its
sacrifices. Like, one could argue that an army is a useless expense
and sacrifice in times of peace, but if you don't work on your
defenses proactively (and GPL is a license that not only respects the
4 freedoms, but also stands up to defend them), you become a sweeter
target, more likely to be defeated and terminated.

--
Alexandre Oliva http://www.lsd.ic.unicamp.br/~oliva/
Free Software Evangelist oliva@{lsd.ic.unicamp.br, gnu.org}
FSFLA Board Member íSÚ Libre! => http://www.fsfla.org/
Red Hat Compiler Engineer aoliva@{redhat.com, gcc.gnu.org}

--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
 
Old 06-19-2008, 04:02 PM
Andrew Haley
 
Default Fedora Freedom and linux-libre

Les Mikesell wrote:
> Andrew Haley wrote:

>>>> Err, one library according to you, was unfree in the sense that
>>>> you weren't allowed to change it in any way; to enhance it, or to
>>>> fix bugs.

>>> That's not an accurate description, although it did have a
>>> restriction on distributing modified versions. I could, of
>>> course, change my own copy and submit bug fixes and enhancements
>>> to the author for incorporation - or make the source modifications
>>> available separately from the package. The restriction was more
>>> about preventing broken versions from being distributed than
>>> enhancements.

>> Sure, but it had that effect, didn't it? If you're not allowed to
>> distribute modified versions without someone else's consent, it's
>> not free (as in freedom) software.

> It's not free the way the GPL redefines the word to mean restricted,
> but it doesn't interfere with your freedom to distribute your
> changes as patches, leaving it clear that it is something different
> from the original author's work that he supports.

I think my meaning was clear. It's not free because you can't
distribute modified versions. And no matter how much you try to
define it away, this basic fact will not change. Yes, you can supply
it with a bunch of patches, but you can't do the obvious thing and
check it in to a public source code control system and work on it
there, since that would mean sharing a modified version. You can't
distribute a modified version as part of, for example, a Linux distro.

It's not free in any sense, except free-as-in-beer.

> In more modern licenses, I'd prefer the ones where you are permitted
> to modify independently and distribute the forked copy if you change
> the package name,

Well, yes. Obviously.

> but it is only in odd circumstances that it even matters or that
> there is any effective difference. Even in GPL circles I think most
> people agree that the best process is to coordinate modifications
> into a single revision tree instead of forking wildly.

Sure, but that's a matter of free choice. That's what freedom means:
you can either fork the software yourself, or you can contribute to
the trunk. The choice is up to *you*. And anyone to whom you give
the software has that same choice, and you can't take that freedom
away from them.

>>> There is the argument that if the author/maintainer stops
>>> updating, the package can die.
>>
>> Quite. And, indeed, that's the inevitable consequence.
>
> It's not at all inevitable since the copyright holder can transfer
> control at any time or might already be a foundation that will outlast
> any possible use for the product.

Sure, or they might not choose to do so. You could say the same about
any software supplied under a restrictive licence.

> But, in technology everyone is better
> off when an old package does die and is replaced by something new and
> improved, and the harm of the GPL is that it's 'work as a whole'
> requirement makes it difficult or impossible for these replacements to
> happen at the component level when the currently best component isn't
> encumbered by the GPL.

Eh? This makes no sense. It's certainly not justified by this
example, anyway.

Andrew.



--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
 
Old 06-19-2008, 04:26 PM
Alexandre Oliva
 
Default Fedora Freedom and linux-libre

On Jun 18, 2008, Les Mikesell <lesmikesell@gmail.com> wrote:

> Alexandre Oliva wrote:
>>
>> Here, take this program here:
>>
>> http://www.lsd.ic.unicamp.br/~oliva/papers/vta/example.c
>>
>> What do you think you're entitled to do with it as it stands?

> Assuming it is permitted for me to get a copy in the first place,

You don't need permission from a copyright holder to receive a copy.
You don't get fined or go to jail if someone sends you spam containing
a copyrighted work, or you pick up a piece of paper with copyrighted
material thrown out of an airplane.

> then I have fair use of it.

Such as...

> I could link my copy with anything else I wanted.

Why do you think you could combine it with other works without my
permission?

> And I could give away my work that called it and some other code.

Depending on how it calls my work, the work you created based on mine
may amount to a derived joint work per copyright law, and then you'd
need my permission to distribute it.

>> Now, if I tell you, I license it to you under the GPL, and you accept
>> it. what you you think you're no longer entitled to do with it?

> Then I couldn't give my additional work away to other people to use
> with their own copies of your work and third party libraries.

No difference.

> They could hire me to re-invent it, but I couldn't distribute my
> work.

If it is your work, you can. If it's our joint work, you'd need my
permission as well, and you'd have permission to distribute it under
the terms of the GPL.

Besides, the GPL doesn't (and couldn't) take away any of your fair use
rights, because it's not a contract (no reciprocation), it's a
unilateral grant of permissions.

>>> No one but the FSF claims that a patch is a derived work

>> How would you defend the claim that something that literally copies
>> portions of a copyrightable work is not?

> It's fair use

Why? How do you (or rather courts) draw the line?

> but I'd settle for being able to do a line-numbered
> edit script that didn't not copy any original material.

Literal copying is not required to create a derived work.

> Or a completely separate work that linked two differently licensed
> libraries that are not included.

Once it links with the other libraries, it contains portions of the
libraries, so it is a joint work.

Just like, it's not because you wrote a novel book that mentions some
famous paintings that you're entitled to print in the book copies of
those paintings, or even vague sketches thereof, and then distribute
copies of the book.

> Yes, we covered the fact that everyone else permits it.

If someone permits it (definitely not everyone), it's presumable that
such a permission is required. Indeed, copyright does require such
permissions. You seem to believe it doesn't, but you're mistaken.

>> Try creating a work based on internal DLLs, for which you have no
>> "developer" permissions.

> Again, I have never heard of any prohibition against sharing such a
> thing with anyone who has their own copy of all required components.

Not knowing or not understanding the law is not an excuse to not
comply with it.

>>> if the resulting 'work as a whole' would be be a derived work,
>>
>> Then it's not "your own original work"

> Define whose work it is, then.

It's a joint work. If you derive a work from mine, in the copyright
sense, then both of us are authors, both of us have a say (veto power)
on how it can be modified and distributed.

> Take the case of an original program
> that links with two other libraries. It is entirely conceivable and
> fairly likly that compatible GPL and non-GPL libraries will be
> available and it is the end users choice at compile/run time. Who
> owns this work?

Whoever owns the source code of the original program and that of the
library whose (minor) portions got copied into the binary.

You defend from a claim of infringement presenting another library and
claiming you derived the distributed executable from it. Then you'd
be bound by the terms of that other library.

In theory, you could clean-room-develop just enough of a duplicate of
the library to be able to create an executable that doesn't copy
anything from the library. It might still be regarded as an
infringement by a court, though, because at times what counts is the
intention, even if the strict letter can be bent to support the facts.

>>> Without the GPL, I can give you a copy of my original work that
>>> links to that library

>> Only if the copyright holder of the library granted you permission to
>> distribute your derived work.

> That's just not true.

I'm afraid you got your facts wrong. Talk to your lawyer about
distributing combined works.

--
Alexandre Oliva http://www.lsd.ic.unicamp.br/~oliva/
Free Software Evangelist oliva@{lsd.ic.unicamp.br, gnu.org}
FSFLA Board Member íSÚ Libre! => http://www.fsfla.org/
Red Hat Compiler Engineer aoliva@{redhat.com, gcc.gnu.org}

--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
 
Old 06-19-2008, 04:37 PM
Andrew Haley
 
Default Fedora Freedom and linux-libre

Alexandre Oliva wrote:
> On Jun 18, 2008, Matthew Saltzman <mjs@clemson.edu> wrote:
>
>> Then can we at least agree that there are sometimes unfortunate
>> consequences to the GPL's failure to permit one to share a work
>> combining two pieces of *free* software because of relatively minor[1]
>> license incompatibilities?
>
> Yeah, it's unfortunate when this happens.

Definitely. It's something that GPL V3 has tried hard to fix,
wherever possible.

However, I must point out that in some cases post-GPL licences
have *deliberately* been worded in a way that makes them incompatible
with GPL code. Whatever the consequences, it's not appropriate to
blame the GPL for those.

>> In fact, I think it's arguable that there are sometimes unfortunate
>> consequences to the GPL's failure to permit one to share a work that
>> makes use of a GPL library and a proprietary library.
>
> Sparing a user from becoming dependent on a piece of proprietary
> software might even be a sacrifice for the user, but it's actually an
> advantage for the user and for society in the long run.

Perhaps.

I think we have to think about, for example, gcc ports. The fact that
people who do ports of gcc are forced to ship the source for their
changes has made a lot of free code available that wouldn't have been
if they had been permitted to link proprietary code into gcc.

Andrew.

--
fedora-devel-list mailing list
fedora-devel-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-devel-list
 

Thread Tools




All times are GMT. The time now is 09:43 AM.

VBulletin, Copyright ©2000 - 2014, Jelsoft Enterprises Ltd.
Content Relevant URLs by vBSEO ©2007, Crawlability, Inc.
Copyright ę2007 - 2008, www.linux-archive.org