On Mon, 22 Sep 2003, Steve Langasek wrote:
>On Mon, Sep 22, 2003 at 11:56:27AM +0200, Josselin Mouette wrote:
>> Le lun 22/09/2003 ? 09:46, Glenn Maynard a ?crit :
>> > On Mon, Sep 22, 2003 at 08:47:26AM +0200, Thomas Hood wrote:
>> > > IBM distributes the Linux driver and the binaries in a tarball that
>> > > it says is licensed under the GPL.
>> > > http://oss.software.ibm.com/acpmodem/
>> > > No source code is provided for the DSP binaries. (N.B., past
>> > > discussions of this issue have reached the conclusion that such
>> > > software can nevertheless be distributed in main.)
>> > If it's licensed under the GPL, and no source is provided, then it can
>> > not be distributed at all, not even in non-free, unless there never was
>> > source to begin with. (I assume this isn't the case, as you said "no
>> > source code is provided", not "no source code exists".)
>> If the binaries were entirely written using assembly code, the binary
>> here equates the source.
>assembly != machine language. If it's written in assembly, the
>source is still assembly, not a binary (despite assembly sometimes
>/appearing/ to be binary gibberish :).
It depends. If there a mutual one-to-one correspondence
between assembler line and DSP processor command it is, mainly, a
differences in format.
You may disassemble binary and pretends that this is a
source code.
