Re: LCC and blobs
Brian Thomas Sniffen writes:
> Matthew Garrett <firstname.lastname@example.org> writes:
> > Brian Thomas Sniffen <email@example.com> wrote:
> >> Michael Poole <firstname.lastname@example.org> writes:
> >>> Lots of people cannot write or modify C code, but we accept as free
> >>> many programs that include C code. The user being inexpert in some
> >>> technique does not render a thing non-free.
> >> But something being *not software* does render it *not software*.
> > dd if=/dev/mem of=/tmp/bios bs=1k skip=960 count=64
> > will happily present you with a copy of your system firmware (assuming
> > you're on x86). If you run ndisasm over it, you'll find it's x86 machine
> > code. You can even extract bits of it and run them. It looks awfully
> > like software. The fact that you lack the skills to turn it into
> > something that you recognise as software doesn't mean it isn't software.
> And putting a paper copy into a good scanner will transform a printout
> of a program into software. The fact that something has a two-way
> conversion to software doesn't make it software. I've asked several
> times for a delineation, in the world where anything convertible to
> software is software, between what is software and what is not. I
> have yet to read such.
Paper is not analogous to a BIOS, and your question is neither useful
nor pertinent. Almost everything in the world can be represented
using some digital representation. However, the DFSG does not mention
"things convertible to software" -- it mentions "software."
A BIOS is normally stored in binary form as executable or
interpretable code plus associated data. Most people would call
executable code in binary form "software;" Debian uses a broader
definition than that. The real question is why you think that
executable binary data is not software.