Why don't I get any credit for completed,uploaded workunits?

Look at Getting started and here

Look at Getting started and here


[edit]
sorry, double-post

1) Credit is only given after THREE different computers have processed the same workunit, uploaded it AND REPORTED the completion, and after the "credit-giver" process (validator/verifier/whatever they call it today) has determined that credit is due. So, it might take quite a while before credit is given. Normally a couple of days; in extreme cases maybe a few weeks. But see next.

2) A recent bug (see news on website front page) means that some complete workunits (no mention of how many) will never get credit. Tough luck all around.

3) They seem to have turned off the "credit-giver" process for the last week or more. Now that they are producing/distributing workunits and accepting uploads again, maybe they'll re-enable that process soon.

read the link below
ttyl
Jeff (Nightowl)
All your answers in one spot:
http://setiweb.ssl.berkeley.edu/transition.php
=====
If you dont like Boinc, then go back to classic seti.

> 1) Credit is only given after THREE different computers have processed the
> same workunit, uploaded it AND REPORTED the completion, and after the
> "credit-giver" process (validator/verifier/whatever they call it today) has
> determined that credit is due. So, it might take quite a while before credit
> is given. Normally a couple of days; in extreme cases maybe a few weeks. But
> see next.
>

I am really disturbed by this new policy. If I volunteer the use of my computer resources, I expect at the very least to get credit and thanks for their use, regardless of whether the results are useful or not. This should not be a function of whether the results agree with other people's (after all it is not my software that is making errors ;-). Perhaps you can keep two different totals, one for work done by each machine and another for results "verified" on multiple machines. The first one would be of interest to me, the second should be of interest to the science team.

> > 1) Credit is only given after THREE different computers have processed
> the
> > same workunit, uploaded it AND REPORTED the completion, and after the
> > "credit-giver" process (validator/verifier/whatever they call it today)
> has
> > determined that credit is due. So, it might take quite a while before
> credit
> > is given. Normally a couple of days; in extreme cases maybe a few weeks.
> But
> > see next.
> >
>
> I am really disturbed by this new policy. If I volunteer the use of my
> computer resources, I expect at the very least to get credit and thanks for
> their use, regardless of whether the results are useful or not. This should
> not be a function of whether the results agree with other people's (after all
> it is not my software that is making errors ;-). Perhaps you can keep two
> different totals, one for work done by each machine and another for results
> "verified" on multiple machines. The first one would be of interest to me, the
> second should be of interest to the science team.
>
This policy is driven by the rampant cheating in S@H1, and some aggressive overclockers that have overclocked so much that the usefulness of all their results are nill.

If your results are always correct, then you should have nothing to worry about. When three results do not agree, then the WU is sent again (and again and again) until there are three results that do agree. These three then get credit. The technique for generating the stats drives behavour. If it makes nosense to cheat or aggressively overclock, then these behavours will be greatly reduced.

>
> I am really disturbed by this new policy. If I volunteer the use of my
> computer resources, I expect at the very least to get credit and thanks for
> their use, regardless of whether the results are useful or not. This should
> not be a function of whether the results agree with other people's (after all
> it is not my software that is making errors ;-). Perhaps you can keep two
> different totals, one for work done by each machine and another for results
> "verified" on multiple machines. The first one would be of interest to me, the
> second should be of interest to the science team.
>
>

I am really disturbed by your attitude. Why would the science team have any interest in keeping track of how many results were returned that were not useful? Why should they waste time tracking useless stats and waste resources storing them? If a machine is not doing usefull work for the project then it does not need to participate. Software does not make the errors, it is hardware that makes the errors. There were far too many instances of cheating in SETI classic by people who were only interested in the stats and not the science. Hopefully BOINC will put an end to the cheating and only get people who are interested in the science. Stats make the project interesting for us but we are in it for the science.

> When three results do not agree, then the WU is sent again (and again
> and again) until there are three results that do agree.

Note that the "agreement" is NOT in the number of credits claimed! If the three don't agree in credits claimed, the "credit giver" process will decide (I don't know the algorithm) how many credits to give all three people. ALL THREE will receive the same amount of credits.

The "agreement" required is in the "science" part of the calculations: number/power/etc of gaussians/triplets/etc. If the calculating clients are not tampered with, then the results will always agree.

The latest NEWS from team may give you an answer...

Greetings from Belgium ;)
Be patient...

> I am really disturbed by your attitude. Why would the science team have any
> interest in keeping track of how many results were returned that were not
> useful? Why should they waste time tracking useless stats and waste resources
> storing them? If a machine is not doing usefull work for the project then it
> does not need to participate.

As I said before, I agree this is what the science team should be interested in.

Software does not make the errors, it is
> hardware that makes the errors.

Not in my present experience - The BOINC client has been unable to download new work units for about 20 hours now - from the messages this is not a hardware or network problem at our end.

There were far too many instances of cheating
> in SETI classic by people who were only interested in the stats and not the
> science. Hopefully BOINC will put an end to the cheating and only get people
> who are interested in the science. Stats make the project interesting for us
> but we are in it for the science.
>
It would be a simple matter to disable any user accounts found cheating.