Francois Gouget fgouget at club-internet.fr
Tue Jun 24 05:41:32 EDT 1997

On Mon, 23 Jun 1997, Tim Charron wrote:
> seems to me that most people are motivated by the stats.  Why not 
> just let the keyservers accept connections from both versions of the 
> clients, but not let information from v1 clients reflect in the 
> stats?  Or, let it all reflect under a common email/host.  People who 

	I don't like this idea of having first-class participants and
second-class participants. It's much like having first-class citizens and
second-class citizens, it create a bad climate.
	Also does this current discussion mean that the users that compile
their client themselves are supposed to  be stuck with old clients,
not as fast with less functionalities ? Will the optimisations of the v2
clients be put in the "v1" clients too ?

	Also a good reason for releasing the source including the servers
source. People are talking here about RC5 and Mersenne but there are other
projects that could benefit from this kind of distributed computing. To
cite those that I know about:

	-- Fermat Numbers Factorisation --

	-- Golomb Rulers --

	-- SETI --

	Should each of these projects redevelop their very own
client-server protocol ? Deschall has developped a client/server protocol
and now that DES is over it has gone to the bin. SolNet has developped one
too but since they still have not released the server sources this effort
is lost too.  This project is developing yet another protocol. Isn't it
time to make things a bit more public and reusable so that all these
projects could essentially share the same code ?

	So as for the current proposisions:

	- bothering the self-compiled client users by every mean available

	  Don't release the source. This is clearer and simpler than
saying "Look, nice source, hey why are you using it!".

	- double checking some percentage of the results

	  But why only the self compiled clients ? If the protocol that RC5
uses is similar to that of SolNet it is easy to reverse-engineered so all
results should be viewed as suspicious. 

	- sending "test" blocks

	  There have been many suggestions concerning this, some of them
very clever. This is certainly worth being investigated. again this should
be applied to all the clients.

	- creating a specific test server/protocol to get blocks for
testing new clients

	  This is also certainly a good idea.

	- automatically downloading new client software

	  Are you trying to recreate ActiveX ?!!! Besides the security
issues, I think that people give their CPU resources *for* a project. This
is also a reason why it would be nice if users just had the choice to give
their CPU time to Mersenne, Seti, Golomb Rulers or whatever other project.
Maybe they're not interested in RC5 after all.

Francois Gouget
fgouget at club-internet.fr                http://www.mygale.org/~fgouget/

To unsubscribe, send email to majordomo at llamas.net with 'unsubscribe rc5' in the body.

More information about the rc5 mailing list