Subject: Out-of-context snippet...
Date: 12 Jun 1999 18:30:07 -0000

...from an email I was writing to somebody else, which some of y'all
might find interesting.

Basically, the person had been commenting on his interest in the
issue of how little we make use of the present power of computers,
vis-a-vis Beowulf and his interest in the Extreme Linux track at
LinuxExpo.  He'd attended our g77 BOF, by the way -- info on which
is at my web site, <>.  He then commented
on the importance of the work of free-software authors like myself,
and the free-software paradigm, when it comes to solving the problem
of making efficient use of spare cycles.  What follows is excerpted
from my response.  It contains some of my thinking in response to the
oft-repeated "but GPL does not make sense for embedded applications"
claim, though not worded as such.


I do think so!  Before I really got into the free-software community, I
was becoming rapidly "stumped" by the problem widespread, efficient
use of "spare cycles" on computers around the world posed, when taking
into account issues of trust (security), robustness (repeatability),
and so on.

The free-software paradigm sure seems to make lots of those problems
go away.  It doesn't necessarily make the *economic* problem go away,
but, given that there are whole bunches of us using otherwise "spare
cycles" to test out g77, gcc, and so on, and share the results, despite
getting no money for that work, it might turn out to not be necessary
to solve the economic problem to get a long way towards the goal.

However, I do think, someday, that a trustworthy system allowing
people and their computers to negotiate payments in exchange for
resources (computing, disk space, backup/archive, etc.), would be
necessary, even if "payments" includes non-monetary transfers.

Such a system need not exclude free software -- in my view, it would
pretty much have to *be* free software (a la GNU) for people to trust
it, since, without the source code, how would people know that their
"agent", a chunk of software, was really negotiating in *their* best
interests, rather than the software's programmers?

(This is generalizable to other rapidly growing technologies.  For
over 20 years to me it's been "obvious" that, for example, the ideal
model -- within human culture, anyway -- for telephony involves
each person employing a trusted agent, a computer program, to dynamically
negotiate among potential carriers for an outgoing call, as well
as with the carrier of an incoming call.  By now, I suspect we've got
the ability, if not the market pressure, to achieve pretty much all
that's necessary within the CPU of the typical digital cell phone.
But who will trust that CPU to always make decisions in favor of the
*user*, rather than the *manufacturer*, of that cell phone?  Nowadays,
my answer is, the best way to determine whether your cell phone is
trustworthy is for its source code, perhaps its software-related schematics,
to be published under the GPL, and for it to allow, even encourage,

        tq vm, (burley)