> Is BOINC responsible for keeping the projects from doing stupid things 
> (unreasonable deadlines, unique credit practices, replication of 1), or 
> should BOINC provide tools and suggest good practices and then leave it to 
> the projects.

I think - the second :)
That's why I strongly resist to consider replication of 1/adaptive 
replication as default.
Defaults are suggestions of good practices from BOINC, right? But this one - 
not good IMO.
Even now some projects (MilkyWay for example) adopted replication of 1 even 
w/o any "adaptive" replication.
And it's their own choice, no problems with that. They know that replication 
of 1 is "deviation" and probably have their own means to deal with it (AFAIK 
MW algorithm resistant to invalid results cause they probably will be killed 
into next genetic algorithm generation, but if SETI's result says there were 
RFI when there is WoW signal we will lost it. Persistence checker probably 
will show "no persistence" and good bye ET :P ).
And I know that "BOINC defaults" tend to be used on SETI.....

_______________________________________________
boinc_dev mailing list
[email protected]
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.

Reply via email to