> If huge effort should be applied, it should
> start from the very root of the problem.
In the past, hardware was either so basic as to not utilize firmware, or,
when it did, onboard processor was barely coping with the job to have spare
cycles for spying around. In that era, "proprietary hardware" meant nothing.
So, only software has to be freed, and that was a manageable task. The whole
suit (OS, DE, utilities) could be dissected and developed by relatively small
teams of people, down to a broke individual.
Good days are over.
Now, hardware is sophisticated enough to have lots of spare cycles for
spying. But how to free the hardware without multi-million dollars of funds?
Design and production of CPU, GPU, networking, etc. components are gigantic
tasks needing multi-millions of dollars. There must be a solution or we can
find ourselves between the rock and a hard place (correct usage of the term,
I hope).
I think the Shakti Project, from this perspective, is of utmost importance. I
regard it as start of a revolution in hardware domain. Luke is also trying to
find a solution for GPU. So far the "solution" seems to be using spare
general purpose CPU cores and running specialized software on it, to turn
that core into a sort of GPU (with too low performance and too high power
consumption). I hope, in future there will be other projects, like Shakti,
focused on the GPU question. And then there is networking hardware...
We are bound by physical constraints in libre hardware production. It's not
like "lets club together and write a module - or whatever". Until the
hardware question is solved, all we can do is "defensive driving", as far as
I can see. In the meantime we can try to prune the free software in regards
to security/privacy issues. I believe you did more than your fair bit in this
regard.
