[vlc-devel] Re: GPL issues ( was: Re: How can add a new codec module to VLC?)

Matthew Romaine Matthew.Romaine at jp.sony.com
Fri May 21 04:15:25 CEST 2004


Derk-Jan

Thanks for sharing the below - I wish I could've been there :)

Gibalou's question caught my attention, and the responses surprised me:

  14:40 < gibalou> considering that only distribution is regulated by 
the GPL, can you actually have code that depends on a proprietary 
library as long as you don't distribute this library with the program ?
14:40 < thedj> yes
14:40 < sam> yes

How important is the ending phrase "...with the program"?  I would have 
thought it to be unnecessary, otherwise someone might interpret the 
above to mean it was okay to write a demuxer and decoder wrapper (much 
like for faad) which would be distributed with the program, and then 
require users to go somewhere else to download a library.  In other 
words, the above exchange might lead one to believe it was okay if 
library B was not distributed literally "with the program" A, but was 
still available somewhere else (albeit without source....)

Perhaps I should clarify my stance here.   Working for a Japanese 
company is not easy; convincing them of the benefits of open source is 
(ironically in some respects if you understand the culture) even harder 
:)  Personally, I would like to make the decoder open source as I feel 
all parties involved would benefit.

Thanks again for an enlightening discussion :)
matt

On 2004/05/20, at 22:15, Derk-Jan Hartman wrote:

>
> On 20 mei 2004, at 13:37, Matthew Romaine wrote:
>> On 2004/05/20, at 15:47, Gildas Bazin wrote:
>>
>>> Last advice, beware that VLC is released under the GPL license so 
>>> you might
>>> have problems with linking to a proprietary library.
>>
>> Heh, I'm glad you brought this up.  I haven't gotten around to 
>> consulting with lawyers yet, but my understanding is that the GPL 
>> doesn't really restrict linking to a proprietary library.  I 
>> understand all modifications I make to the VLC code base -- including 
>> file parsing and function calls to the library, etc. -- will have to 
>> be released, but aside from the concern that people will know what 
>> the function names are inside the library, is there anything else 
>> that I should note?  The GPL doesn't require that I also release the 
>> proprietary library, right? (i.e., my contributions for the time 
>> being may be unusable...)
>
> Welcome to one of the great debates on the interpretation of the GPL 
> license. Above explanation is the one used by many. VideoLAN however 
> follows the GPL to the letter ( mplayer and Xine dont ) and the GPL 
> strictly forbids linking of GPL programs to proprietary, closed code, 
> unless they are part of the OS itself. A sigma proprietary library can 
> not be considered to be part of the OS and therefore a GPL'ed program 
> could not make use of it. This is how the GPL works in our reading of 
> it, and in our eyes mplayer, Xine and a lot of other projects violate 
> their own license...
>
> As a show of some of the discussion this results in:
>
> 14:31 < thedj> mm, GPL and proprietary lib. issues. don't you just 
> love them?
> 14:31 < bigben_> well, I thought GPL was quite clear about it...
> 14:31 < thedj> actually it isn't
> 14:32 < thedj> it talks about derivative works. but it's hard to 
> define what that is.
> 14:32 < bigben_> so, let's at least tell him the way we understand 
> it...
> 14:32 < thedj> for instance this section
> 14:32 < thedj> If
> 14:32 < thedj> identifiable sections of that work are not derived from 
> the Program,
> 14:32 < thedj> and can be reasonably considered independent and 
> separate works in
> 14:32 < thedj> themselves, then this License, and its terms, do not 
> apply to those
> 14:32 < thedj> sections when you distribute them as separate works.
> 14:33 < sam> well, a program that uses a given library can obviously 
> not be considered "independent"
> 14:34 < thedj> sam: but the two sourcecodes can be independently 
> distributed, as long as you don't compile them....
> 14:34 < thedj> rather a weird situation :)
> 14:34 < sam> thedj: this is circumvention
> 14:34 -!- Psycho[B]ud [~PsychoBud at h000102c59fbb.ne.client2.attbi.com] 
> has quit [Read error: 110 (Connection timed out)]
> 14:34 < thedj> no, it's what it says...
> 14:35 < sam> who, it?
> 14:35 < thedj> sam: GPL.
> 14:35 < sam> if program A uses library B, it is obviously derived from 
> B; you can easily prove that by stating that without B, A would not 
> exist
> 14:35 < thedj> if the prop lib is written before the program, you 
> cannot call it a derivative work.
> 14:36 < Defaf> when is 0.7.2 coming out?
> 14:36 < sam> I am not stating that the library is a derivating work of 
> the program
> 14:36 < sam> I am stating that the program is a derivative work of the 
> library
> 14:36 < sam> s/ing/ive/
> 14:36 < thedj> true.
> 14:38 < thedj> still derivative is not a good description.
> 14:38 < sam> derivative is a legal term
> 14:38 < thedj> which will never hold up in court ....
> 14:38 < thedj> a 'user' would be more correct.
> 14:39 < sam> user does not mean anything
> 14:39 < sam> derivative has a legal meaning and it has been used again 
> and again in court; not related to the GPL though
> 14:40 < gibalou> considering that only distribution is regulated by 
> the GPL, can you actually have code that depends on a proprietary 
> library as long as you don't distribute this library with the program 
> ?
> 14:40 < thedj> yes
> 14:40 < sam> yes
> 14:40 < thedj> by the letter of the GPL, i think you can
> 14:40 < thedj> actually you could even use it yourself.
> 14:41 < thedj> as long as you don't distribute
> 14:41 < gibalou> and users could install the library themself as long 
> as they know how do find/do it.
> 14:42 < thedj> derivative is incorrect naming. it implies being based 
> on. using as a basis, using as a source.
> 14:42 < gibalou> but I guess having instructions on how to do this 
> would not be legal
> 14:42 < thedj> VLC is not the based on a proprietary lib
> 14:42 < thedj> to derive actually means: "to trace the origin of 
> something"
> 14:43 < sam> "derivative work" does not mean that
> 14:43 < gibalou> if you use a feature from a library then your code is 
> "based" on that feature.
> 14:43 < thedj> this is why derivative was a bad choice. it implies 
> origins.
> 14:43 < thedj> sam: yes it does. it means work based on another 
> original work
> 14:43 < sam> OR "contains parts of"
> 14:45 < sam> you can bypass any legalese by nitpicking ad absurdum; 
> however you can never bypass the author's will when there is a 
> misunderstanding about the meaning of a sentence in the GPL
> 14:46 < sam> thedj: when you add support for library B in program A, 
> you are creating a derivative work of both A and B; since the code 
> gets reinjected into A we keep calling it A, but in essence the 
> derivative work is not A, it is "A plus the addition of this 
> particular piece of code"
> 14:46 < thedj> i disagree.
> 14:46 < thedj> derivative work does not mean.. contains parts of.
> 14:46 < thedj> not according the dictionaries i have here...
> 14:47 < thedj> trough process of derivation
> 14:47 < thedj> by deriving something
> 14:47 < thedj> that's what is says here.
> 14:47 < sam> does it say anything about the legal meaning?
> 14:47 < thedj> and "to derive" is clearly defined with the words 
> origin of development.
> 14:47 < ozone> nothing quite like licensing arguments
> 14:47 < thedj> sam: no.14:47 < thedj> ozone: :)
> 14:48 < ozone> i think dynamically loading plugins is rather murky 
> ground
> 14:48 < sam> thedj: I suggest you open *any* legal dictionary and look 
> into it
> 14:48 < sam> http://www.legal-definitions.com/derivative-work.htm14:48 
> < sam> http://www.legal-database.com/derivative-work.htm
> 14:48 < ozone> best to consult a lawyer to find out what to do ... and 
> until then, any debate is somewhat irrelevant
> 14:49 < sam> ozone: this has been discussed for ages by countless law 
> advisors for the FSF and other fundations
> 14:49 < sam> it works as designed
> 14:49 < gibalou> http://www.linuxjournal.com/article.php?sid=6366
> 14:50 < sam> the whole KDE fiasco is based exactly on that (and was 
> the reason for Qt becoming GPL)
> 14:50 < ozone> sam: for all countries?
> 14:51 < sam> ozone: no idea. but copyright law is rather homogeneic 
> accross countries
> 14:51 < thedj> sam: "derivative work" : For copyright purposes, a new 
> work based upon an original work to which enough original creative 
> work has been added so that the new work represents an original work 
> of authorship
> 14:52 < thedj> even the law dicts don't agree :)
> 14:52 < ozone> sam: i don't know enough about to say anything 
> authoratative
> 14:52 < thedj> derivative work
> 14:52 < thedj> : a piece of intellectual property that substantially 
> derives from an underlying work
> 14:52 < sam> they still all agree on implying being based on
> 14:52 < ozone> but from internal legal counsel that i've heard at 
> work, they don't want to go near issues of dynamic loading
> 14:52 < thedj> yes based on, but not contain parts of.
> 14:52 < ozone> YMMV, as always
> 14:53 < sam> thedj: if it didn't contain parts of, it wouldn
> 14:53 < sam> 't be a derivative work
> 14:53 < thedj> and that latter definition was from webster dictionary 
> of law
> 14:53 < sam> I don't see your point
> 14:53 < thedj> based on != contains parts of
> 14:54 < ozone> anyhoo, have fun arguing guys :)  -> bed for me!
> 14:54 < thedj> hehe
> 14:54 < sam> thedj: the latter implies the former; this is more than 
> enough for my point
> 14:55 < thedj> if something is based on something else, it by 
> defintion contains parts of.
> 14:55 < thedj> but if something contains a part of something its not 
> by definition based on that.
> 14:55 -!- Bond128 [~88d7fbb3 at krishna.via.ecp.fr] has joined #videolan
> 14:56 < thedj> i like these discussions :)
> 14:56 < sam> you got it mixed up
> 14:56 < sam> if something contains a part of something it is obviously 
> based on it!
> 14:56 < sam> unless they are sufficiently independent, which they 
> aren't here
> 14:57 < thedj> if my TV contains a button, then my TV is not based on 
> buttons. it's based on a katheder diode.
> 14:58 < sam> please restrict the discussion to software or at least IP
> 14:58 -!- xxcv [asdf at 203-213-81-61-syd-ts13-2600.tpgi.com.au] has quit 
> [Read error: 60 (Operation timed out)]
> 14:58 < sam> the conccept of a TV is based (amongst others) on the 
> concept of a katheder diode.
> 14:58 < sam> same for buttons
> 14:59 -!- Bond128 [~88d7fbb3 at krishna.via.ecp.fr] has quit [Client Quit]
> 15:00 < thedj> yet is your the concept of a TV anything less without a 
> button? would you say the concept of TV is derived from the concept of 
> buttons.
> 15:02 < sam> without a button it would be a different TV
> 15:02 < thedj> you use buttons to control the TV, but does that make 
> the button a part of the concept of a TV.
> 15:02 < gibalou> maybe not the concept but surely part of the TV ;)
> 15:02 < sam> if you insist that a TV is not based on the concept of a 
> button, you need to call your TV that has buttons "a TV with buttons"
> 15:03 < gibalou> I guess we can say bye-bye to the atrac decoder ;)
> 15:03 < thedj> gibalou: derived from vs. part of :)
> 15:04 < thedj> sam: and i also buy a TV with teletext
> 15:04 < gibalou> personnaly I agree with sam. The GPL was designed to 
> keep the source code free.
> 15:04 < thedj> or even better a TV with color.
> 15:04 < gibalou> allowing the use of proprietary libaries would remove 
> a big part of that benefit.
> 15:04 < thedj> gibalou: i do to :)
> 15:05 < Defaf> when is 0.7.2 due?
> 15:05 < Defaf> i mean
> 15:05 < thedj> gibalou: still, from a legal point of view i think you 
> could actually make a reasonable case of it....
>
> However as sam said:
> you can bypass any legalese by nitpicking ad absurdum; however you can 
> never bypass the author's will when there is a misunderstanding about 
> the meaning of a sentence in the GPL.
>
> And the opinion of the Authors of VideoLAN has always been very clear. 
> A completely Free and Open Source program in it's strictest meaning.
>
> DJ
> ---
> Universiteit Twente
> Derk-Jan Hartman (d.hartman at student.utwente dot nl)
> http://home.student.utwente.nl/d.hartman
>
> -- 
> This is the vlc-devel mailing-list, see http://www.videolan.org/vlc/
> To unsubscribe, please read http://developers.videolan.org/lists.html
> If you are in trouble, please contact <postmaster at videolan.org>
>
>
----------------------------------------
Matthew M.I. Romaine
IMNC Audio Codec Group
Sony Corporation
phone: 03-5448-6065
fax: 03-5448-5617
----------------------------------------

-- 
This is the vlc-devel mailing-list, see http://www.videolan.org/vlc/
To unsubscribe, please read http://developers.videolan.org/lists.html
If you are in trouble, please contact <postmaster at videolan.org>



More information about the vlc-devel mailing list