[RC5] OGR percentage done - WHY impossible?
Zorba the Hutt
zorbathut at uswest.net
Tue Mar 20 19:34:26 EST 2001
> At a minimum, you will get full credit in stats.
Someone else mentioned this also, but I feel it bears repeating - full
credit in stats? Whoopee. Just what I wanted. Stats. I really don't care
about stats - it's somewhat nice to be able to say "I'm #15000" (or whatever
it is that I am) but I'd rather know that my CPU cycles are going towards
something useful and not to the great bit bucket in the sky.
> This discarding of work is nothing new for d.net. We have had to
> invalidate data from other projects in the past, when a particular client
> build was causing invalid results. In every case, we attempt to give full
> credit in stats to the submitter if the fault was not theirs.
As I said, I'm not concerned with losing stat placement - I'm more concerned
with losing work. And I'm especially concerned with saying "This discarding
of work is nothing new for d.net".
And I'm extremely concerned with the fact that D.NET DIDN'T TELL US. We had
to spend days (weeks?) griping on the mailing list - and were finally
informed by someone *else*, only somewhat connected to the project, at which
point dnet says "well . . . yes . . . we *might* have a mistake. You *might*
be losing your blocks. But don't worry! You'll get stats credit!"
This is why I'm not using distributed.net on my main computer anymore. I
don't care about stats credit, I want to be doing something useful. And
wasting CPU cycles because the coordinators don't want to tell us their code
is buggy is not useful. If you make a mistake, just 'fess up - it's better
than dragging it out. I'd much rather hear "We're not sure this code is
right - we're checking it" followed by "oops, we made a mistake - the code
*does* work. Sorry for the false alarm!" than this.
To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest
More information about the rc5