Re: Heart of the debate
Raul Miller wrote:
> > On Tue, Feb 15, 2000 at 12:02:31PM -0500, Andreas Pour wrote:
> > > I think this is where you went off-track. Section 2 only refers to
> > > source code distributions (as it requires the modifications to be
> > > distributed under Section 1 and Section 1 deals only with source
> > > code).
> >
> > I disagree.
> >
> > Section 0 includes the text:
> > <quote>
> > (Hereinafter, translation is included without limitation in
> > the term "modification".)
> > </quote>
> >
> > Section 2 begins:
> > <quote>
> > 2. You may modify your copy or copies of the Program or any portion
> > of it, thus forming a work based on the Program, and copy and
> > distribute such modifications or work under the terms of Section 1
> > above, provided ...
> > </quote>
> >
> > Since executables are a translation of the source, the provisions of
> > section 2 also apply to executables.
On Tue, Feb 15, 2000 at 03:22:35PM -0500, Andreas Pour wrote:
> It's obvious that a binary is a derived work of the Program, but that
> does not mean that every single provision of the GPL applies to it.
> You can only apply what makes sense to apply.
You're implying that not every single provision of the GPL applies to
the distribution of a binary. I've yet to see any legal foundation for
this belief.
> Section 2 requires distribution under Section 1 -- which explicitly
> refers only to source code. If you can explain to me how you can
> distribute a binary under Section 1, then we can explore the
> conditions to such distribution (found in 2(a) - 2(c)).
I already explained this (see material quoted at the top of this message):
the binary is a translation of the source code -- it's the source code in
a specialized form.
> Note that Section 1 specifically says:
>
> You may copy and distribute verbatim copies of the
> Program's *source code* as you receive it.
>
> It could have left out "source code" if it was to mean what you say it
> means. But since those words are there, we must give them meaning.
> Since w/out the terms "source code" there, a binary might be covered by
> Section 1, one can only conclude that a binary does not fall under
> Section 1.
You don't get to ignore sections of the license because you don't care
for the way that they're phrased. It's very clear, from Sections 0 and
1 of the GPL that distributing a binary under Section 1 of the GPL means
distributing a verbatim copy.
You might as well argue that since Section 1 talks about distributing
source code "as you receive it" that it's nonsense to talk about
distributing modified copies of the source code under the terms of
Section 1. Yet, this is exactly what Section 2 allows you to do.
Furthermore, Section 3 talks very specifically about distributing the
program in object code or executable form under the terms of sections
1 and 2.
Finally, Section 5 states that you are prohibited from distributing
the program (or any work based on the program) if you don't accept the
license. So if, as you state, you can't distribute the binary under
the terms of section 1, then you have no legal right to distribute that
binary at all.
> Since the conditions to Section 2 apply only to distributions of
> modified works under Section 1, and since Section 1 does not apply to
> binaries, the conditions of Section 2 do not apply to binaries.
>
> Isn't logic wonderful?
Fabulous.
> Incidentally, it is my reading that translation does not refer to
> compilations into binary. I think it refers to translation into
> another human-readable language -- like translating C++ into C, or C
> into Perl, or C into assembler.
So now you claim that whether or not it's a translation depends on the
target language? What legal reason do you have for this?
> That way use of the terms makes sense -- Section 3
> deals with binaries and Section 2 with source code (including
> translations into other languages). Also, it is not common usage to
> refer to a binary as a "language" -- rather, a binary is referred to as
> a set of"instructions". Binary is really not a language -- nobody can
> read it :-). For dictionaries noting the difference b/w computer
> languages and machine code, see, e.g.,
> http://www.instantweb.com/~foldoc/foldoc.cgi?machine+code and
> http://www.instantweb.com/~foldoc/foldoc.cgi?programming+language.
If this belief of yours was legally sound then binaries would receive
no copyright protection whatsoever. Just build a new executable
and voila.
> Incidentally, if notwithstanding the very clear language you believe
> that Section 2 does apply to a binary, Section 2(a) says:
>
> (a) You must cause the modified files to carry prominent
> notices stating that you changed the files and the date
> of any change
>
> If you can explain to me how Debian does that with the binaries it
> distributes, I would consider myself enlightened.
Well, of course every file has a timestamp on it indicating when it
was built. That's built into the file system.
--
Raul
Reply to: