no blanket firmware exception, please (was: Re: Proposed wording for the SC modification)
- To: email@example.com
- Subject: no blanket firmware exception, please (was: Re: Proposed wording for the SC modification)
- From: Toni Mueller <firstname.lastname@example.org>
- Date: Wed, 24 Dec 2008 00:13:40 +0100
- Message-id: <email@example.com>
- In-reply-to: <firstname.lastname@example.org>
- References: <20081112111410.GW5485@anguilla.noreply.org> <20081114201224.GA11008@intrepid.palfrader.org> <email@example.com> <1226914358.4295.23.camel@shizuru> <20081117130540.GE8151@anguilla.noreply.org> <20081117131949.GB7545@kunpuu.plessy.org> <firstname.lastname@example.org>
On Mon, 17.11.2008 at 09:38:19 -0600, Manoj Srivastava <email@example.com> wrote:
> On Mon, Nov 17 2008, Charles Plessy wrote:
> > Can the Secretary clarify again what will hapen if Peter's option is voted ?
> That GR clearly refines the DFSG statement that all programs
> need source code. This supersedes the current DFSG, which allows for no
> such exception. So the we need to amend the FSG wiht the changes after
> the 3:1 vote. (Aside, on a personal note, anything else, to me, smells
> of deceptive and underhanded handling of our social contract).
without repeating the suggested new wording for the DFSG, I'd like to
note that the concept is imho technically flawed, and serves to corrupt
our freedom and the integrity of our systems in most likely unintended
ways as a side effect.
This is because it's becoming increasingly difficult to draw the line
between "firmware", and a whole operating system for a co-processor.
Think of devices that increasingly take control of our computers, like
GPUs. nVidia's number-crunching graphics cards, or (intel) network
cards that autonomously write stuff into the CPU's cache, even, spring
to mind as examples of what's happening already today.
If this tend were to continue, and I have very little doubt about that,
we will end up with a "small" CPU that will run on free software
because of some legacy decisions by the chipmakers, but all interesting
stuff will happen on powerful co-processors, formerly "peripherals",
which are exempted from running code that we have source for, and which
will therefore be quite opaque to us.
My fear is that handing out a blanket waiver for source code for such
devices will get us into deep trouble before long, and therefore reject
this move. I think that this issue most likely needs further
discussion, and maybe even decisions on a case-by-case basis.
[ OBS: Yes, I know that the vote is already over, but I missed these
points in the discussion so far (but without having waded
through all of it) and thought that these problems need to be