Re: on firmware (possible proposal)
----- "Frans Pop" <firstname.lastname@example.org> wrote:
> And I believe that Debian is becoming increasingly marginal because users
> are driven away to other distros. Sure, it is nice that a lot of those
> users go to derived distros instead of "real" competitors, but IMO it is
> still unhealthy if Debian's own user base becomes too small. After all,
> we largely depend on having that user base for a healthy turnover in DDs
> and having motivated translators and such.
If we handle things properly, it is like the GSM standards body worrying that it doesn't actually build phones. Ubuntu has helped us suck the air out of Red Hat and Novell in a way that no one would have thought possible. Now, to your point, I would like to see them be a little more direct about co-marketing with us and it seems that they are coming around to that way of thinking as time goes on. I think they've eventually come around to the idea that "Debian done right" doesn't fly nearly as well as "Debian + Added Value". Being derogatory about the foundation of your product is backwards and nonsensical.
What we need to do is get these distributions to view co-marketing with Debian as a positive thing, like the "Real" Milk products or the Java (gah! dare I say it!) brand. To me, real Debian derived distros would provide users with the core Debian value of "choice" and then add value on top of that. What we need to do is make some formal rules about how "choice" is provided. For instance, a Debian based platform that locks the user out of root and has cryptographically secured firmware would not meet the criteria of an "official derived distro".
The kind of protections I'm outlining here will become far more important than the bits of firmware in the main distribution as time goes on. If, in the future, there are no computers to install Debian on (ie. everything becomes more embedded and network based) then it won't matter a bit what is on our CDs. Engaging the platform manufacturers and other value-added resellers is going to become ever more critical.
> If we are selective in our concessions to pragmatism and careful in the
> way we implement things (which always has been one of the strengths of
> Debian), I personally don't see the problem.
> I'm well aware that what I tend to think of as the "DFSG hardliners" in
> the project do not like this, but why not discuss things openly (and
> hopefully without too much flaming) and vote on it? Let's see what the
> project as a whole thinks.
I'm not disagreeing with you on the fact that bundles of proprietary bits are going to become harder to get away from as we move towards more embedded looking platforms. I'm just saying that distributing those bits through the core development process is a mistake and that value-added distributors are the way to go. That's what we're already doing and it works. What we should do is increase the effectiveness of our current process by making it more formal and explicit.
Martin Michlmayr actually proposed some similar ideas under the name "Debian Labs", which I immediately reserved as a means of making a point about trademark infringement. That URL could still be a good place to do this kind of work and I'm happy to hand it back to SPI if someone is serious about doing that kind of work on it.
Ean Schuessler, CTO Brainfood.com
email@example.com - http://www.brainfood.com - 214-720-0700 x 315