[HARDWARE] Mac Questions -

John L. Bass jbass at dmsd.com
Tue Mar 12 22:58:47 EST 2002


	Distributed projects like RC5 are done

	- by people who: a) love stats b) love money c) are doing it for fun
	  (because it's their hobby)
	- isn't valuable only to some people, not the whole society
	- has a predefined end

	Distributed projects like FightAids at Home are done

	- by people who: a) love participating in a 'social' project or c) [above]
	- MAY be valuable to all people
	- has not a predefined end
	- is ran by a for-profit org that could easily decide that the project
	  is not longer worth it, and switch to another one

Obviously the two are not mutually exclusive, or even this well defined. Guidance
on how to construct a distributed "public good" project which statisfies the reward,
status and competitive desires sounds critical for a "sucessful" design. Clearly
the FightAids at Home guys made some clear mistakes, and public researchers not tied
to corporate dollars probably need to understand what the Dnet community needs in
designing a distributed application that would be sucessful in this environment.
Your views are probably valuable input to some researcher in search of cycles, so
thanks for that valuable contribution.

I think we as a Dnet community need to reframe Dnet as a general resource, not a
crypto bench, and I expect there are some grant funded projects which have many
tens or hundreds of thousands of dollars that could be pumped into Dnet as lotto
prizes for participating in crunching their numbers. There doesn't need to be
a completion point for them to pay off - just dribble money into Dnet at some
fraction of what they would have to pay for supercomputer time - and randomly
select participants all thru the project with a "win".

I don't think we need to hand free cycles to any big corporate effort - crypto or
medical - but reach out to help smaller researchers that simply do not have access
to corporate supercomputers to get their work done.

	> Dnet service do nothing but Dnet for the purpose of building personal
	> stats.

	Those 'cycles' are contributed to the project whether you run it solely for
	this purpose or not. Let's say I have 100 machines (that I do not have).
	Let's say I'm running those machines solely for dnet. These 100 machines
	could do around 300 Mkeys/s or even 400 Mkeys/s.
	This obviously would help the project AND it helps it to finish faster(!) = to
	save cycles = to save electricity and money.

It doesn't matter if the cycles/energy are done today, or tomarrow, if they are used
there is no savings.

John
--
To unsubscribe, send 'unsubscribe hardware' to majordomo at lists.distributed.net



More information about the Hardware mailing list