Re: Let's stop feeding the NVidia cuckoo

[Matthew Garrett <mgarrett@chiark.greenend.org.uk>]

Francesco Poli <frx@firenze.linux.it> wrote:
> On Mon, 28 Feb 2005 10:16:46 +0000 Matthew Garrett wrote:
> 
>> If we actually upheld this standard at present, it would result in us
>> removing a large number of packages from Debian.
> 
> I think that these issues are sarge-ignore because of GR2004-004, but
> will be release-critical bugs post-Sarge.

That's, uh, entirely insane.

>> If a JPEG can be considered "free enough" under some circumstances,
>> I'm confused as to why it's not always good enough.
> 
> OK, think of a program.
> I give you a file written in C, that can be compiled by gcc into the
> binary executable.
> 
> Am I giving you the source code?
> Yes, in most cases, I am.

Indeed.

> But what if the program is a parser generated by Bison?
> Now the C code is not source code anymore.
> The grammar description is the real source code.

Also true.

> If C code can be considered "free enough" under some circumstances, why
> is it not always good enough?
> Because it's not always the "preferred form for modification", that's
> why!

No. Autogenerated C is not the preferred form for modification, but nor
is it a practical form for modification (in most cases - this is not
always true). However, in almost all cases it *is* practical to modify a
JPEG.

Machine-generated C code is fundamentally different to hand-written C
code, in the same way that a machine-generated JPEG (for instance,
something designed to stress test JPEG decoder algorithms) is
fundamentally different to the more common sort. There is no fundamental
difference between a JPEG that was derived from a lossless format and
one that has always been a JPEG.

-- 
Matthew Garrett | mjg59-chiark.mail.debian.legal@srcf.ucam.org