[RC5] Possible V3 addition.

Mark Salyer marks at foothills.net
Thu Mar 12 10:54:19 EST 1998


Well, I never thought that most of the replies would be *this* negative, I
figured we could debate the good and bad instead of just bashing to begin
with, but I won't take anything personally :).

>Mark Salyer wrote:
>>
>> I was wondering what other people thought about a couple of ideas that I
had
>> concearning the rc5 contests.  Some people have had some concearns lately
>> about getting their bosses approval on running the Distributed.net
clients.
>
>This is a major point of contention in some places, personally
>I think that we should expend more effort getting more individuals
>to join the resource pool, and worry about recruiting corporate
>cpu farms later.
>

I do not think that we should just forget about those corporate cpu farms,
because that is where the real power is..  Yes there are millions of
individuals that are out there, but a few larger coporations could easily
top that..  We want to prove that Rc5 is not good enough, the corporations
have just as much if not more at stake than the average individual...

>> One of the ideas that I had, that would really help me and the company I
>> work for out tremendously, would be to add some "time correcting" code to
>> the current clients.  Allow the proxies to send out a current time when
>> blocks are distributed, and then let the clients change the time of the
>> computers when they retrieve new blocks.  This would only be done when
>> fetching/flushing is going on, so it should not slow down the clients
that
>> much..  This way, we as sysadmins are getting extra value out of the
>
>"We as sysadmins"??? What kind of systems do you administer that
>do not support the network time protocol.? What kind of systems
>would let a user level process set the system clock?

I administer systems that could use their time correct..  There are not alot
of *free* utilities out there that do this, and when you are a non-profit
yourself, you can't afford $10 per machine by 2000 machines to just set the
time..  You are more concearned about the *individual*, and that is fine,
they could use an ini setting to turn this off..  Some corporations could
use this and I am speaking for one..  I would as soon trust DCI as any other
small programming shop to set the time on my servers..

>
>w95 and nt already have various third party programs for access
>to the the USNO clocks or something equivalent. *nix systems have
>a number of network time protocol support methods basically built
>in and I for one would not let the rc5 client fiddle with the
>system clock on my linux box. It runs as the lowest priority
>process and has no write privileges to anything but its own
>files.


Like I said, that is fine, if you are fortunate enough to be running an
entire enterprise on Linux boxes, great..  If not, and you do what the rest
of us does, you have to depend on 3rd party applications, and I would rather
have one program doing both from one source that I trust, than to run
several products doing the same thing..  Slows everything down (not greatly,
but two programs in memory is larger than one), and adds traffic to network.
Besides, why should I trust the Linux developers more than the DCI
developers?  And I must say, I have been running Linux since 92...


>Having the d.n clients do "something else that might be considered
>constructive" might be a decent idea, but not this specific one.

I would not think you could find something as quick to do that would be
somewhat useful.  Remember, though it may be free for Unix, It is not
necessarily the same for Win/Macs.

>We actually have already made use of the rc5 client-perproxy-proxy
>communication chain as a crude network monitor. The perproxy and
>its logs can give you a quick and easy snapshot of which boxes
>are up. Take a look at some of the pproxy log->stats processing
>programs and see what they produce. I think that you'll find most
>of these suggestions have actually already been implemented.
>

Yes, it is somewhat crude, but it could use some work..  What I was trying
to say, is that with some small modifications, (Remember, I said this may be
a seperate effort) could turn into something more useful.  One this project
(Stats/Monitor), I was hinting at another group *loosely* tied to DCI to do
this..  There are several posts of people wanting to help that never get
answered.   The web page is slowly updated, and several people have offered
to help.  (I know that DCI has their reasons for not expanding the core
group to quickly, that is not the point here) I am just saying, that there
would be plenty of help..

You and I are on opposite ends of the argument here, you are more thinking
of the individual, and I am more thinking of the company.  When I go to the
head of the company and say there is something that I would like to run on
non-critical machines, I would like to be able to say that this does
something for us as well.  Non-Computer people sometimes do not care about
encryption, and I thought that would help.. I just wanted to know what
everyone else thought, and I think I now know that no one thinks it was a
good idea..   If I have offended someone, then I apologise...

Marks


>--
>Marc Sissom                       http://www.dnaent.com
>Design Engineer                     voice: 972/644-3301
>DNA Enterprises, Inc.                 fax: 972/644-6338
>--
>To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
>rc5-digest subscribers replace rc5 with rc5-digest
--
To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest



More information about the rc5 mailing list