I am suspending E@H and may detach. The employment of Akos as a consultant seems to have been completely mishandled. We have an only half-optimised stock app and another round of chaos and insurrection - exactly what I thought 'they' were trying to avoid.
Well, that may be a little harsh. :-)
I wonder if UMW has the resources to spare to start a beta project like SAH does for pushing the envelope. ;-)
Alinator
That would be a very good idea. I'm already taking part in a few alpha and beta projects .
Also would like to point out that I have been careful with the testing to make sure that i was not submitting garbage to the project. I did not even mix apps while doing results because I don't think that is a proper way to test. I did complete runs with one test app and verified that is was valid b4 allowing anymore to be done with that app because I did not want to see us get in this kind of situation. I guess there are a lot of people that are not as careful when they test. So once again a few spoil it for the rest.
:(
That would be a very good idea. I'm already taking part in a few alpha and beta projects .
Also would like to point out that I have been careful with the testing to make sure that i was not submitting garbage to the project. I did not even mix apps while doing results because I don't think that is a proper way to test. I did complete runs with one test app and verified that is was valid b4 allowing anymore to be done with that app because I did not want to see us get in this kind of situation. I guess there are a lot of people that are not as careful when they test. So once again a few spoil it for the rest.
:(
Didn't mean to imply you were being careless, I know you're no "nugget" here in EAH. :-)
I guess the main reason is the rate they were being developed and the reduction in the initial replication made it more likely this could occur.
In any event, one would like to think the saviness level of EAH members is higher, so I don't think this will lead to any kind "flap" like even the smallest change to anything at SAH does frequently. :-)
Please, don't use any patched versions of Einstein@Home app.
The bad results caused some problems in the database.
Thanks!
edit: Probably these units will be send out again. ( about 5000 WUs at moment )
Hi Akos,
I don't know if that is the best way. Same of your optimalization produce invalid results, thats true. But it seem that other optimalization produce correct results and faster as official application.
I'm not sure if all people are so civic-minded and go back to the official application. So I think that Bernard in cooperation with you can implement your right improvement into the next official application and for transition period say which optimalization are correct and which are uncorrect. Second way should be increase quorum on 3 and send at least one of this to computer with nonWindows OS.
What does this mean for us that have edited our client_state files because of the 4.10 mess? Will just running out my cache be enought?
Unfortunatly means i'll have to wait 8hours on this machine and probly 18hours on second.
also all the versions producing bad results and having bugs (like the 0713) could be removed everywhere and only the "good" versions like 0007, 0709 and 0712 could be distributed... Dont see where then would be a problem. The 0712 seems to be very good.........
Maybe this would be a good moment to try to implement a project setting that says "I want to receive Test Workunits and Applications", something like "inline beta testing".
For each pair of stock WUs they could deliver one beta WU which could be validated and credited like the stock WUs - but without influencing the scientific results, those should only be taken from the two stock results.
Pirates doesn't have WUs at
)
Pirates doesn't have WUs at the moment, maybe they could help out for beta testing
RE: RE: I am suspending
)
That would be a very good idea. I'm already taking part in a few alpha and beta projects .
Also would like to point out that I have been careful with the testing to make sure that i was not submitting garbage to the project. I did not even mix apps while doing results because I don't think that is a proper way to test. I did complete runs with one test app and verified that is was valid b4 allowing anymore to be done with that app because I did not want to see us get in this kind of situation. I guess there are a lot of people that are not as careful when they test. So once again a few spoil it for the rest.
:(
98SE XP2500+ @ 2.1 GHz Boinc v5.8.8
@ Ananas Hey, now there's
)
@ Ananas
Hey, now there's a thought!! :-)
Alinator
um ... incomplete idea though
)
um ... incomplete idea though - I forgot the Arrrrrr!
If the 'problem' is the
)
If the 'problem' is the version 4.10 results, I'm going to be really p*ssed!
Dead men don't get the baby washed. HTH
RE: That would be a very
)
Didn't mean to imply you were being careless, I know you're no "nugget" here in EAH. :-)
I guess the main reason is the rate they were being developed and the reduction in the initial replication made it more likely this could occur.
In any event, one would like to think the saviness level of EAH members is higher, so I don't think this will lead to any kind "flap" like even the smallest change to anything at SAH does frequently. :-)
Alinator
RE: Hi people! Please,
)
Hi Akos,
I don't know if that is the best way. Same of your optimalization produce invalid results, thats true. But it seem that other optimalization produce correct results and faster as official application.
I'm not sure if all people are so civic-minded and go back to the official application. So I think that Bernard in cooperation with you can implement your right improvement into the next official application and for transition period say which optimalization are correct and which are uncorrect. Second way should be increase quorum on 3 and send at least one of this to computer with nonWindows OS.
What does this mean for us
)
What does this mean for us that have edited our client_state files because of the 4.10 mess? Will just running out my cache be enought?
Unfortunatly means i'll have to wait 8hours on this machine and probly 18hours on second.
also all the versions
)
also all the versions producing bad results and having bugs (like the 0713) could be removed everywhere and only the "good" versions like 0007, 0709 and 0712 could be distributed... Dont see where then would be a problem. The 0712 seems to be very good.........
An option over a separate
)
An option over a separate beta project :
Maybe this would be a good moment to try to implement a project setting that says "I want to receive Test Workunits and Applications", something like "inline beta testing".
For each pair of stock WUs they could deliver one beta WU which could be validated and credited like the stock WUs - but without influencing the scientific results, those should only be taken from the two stock results.
(not a new idea btw.)