Hi
Copyed from D20L´s forum. 3 postings.
_________________________________
the stats contain the true node stats. Node XYZ would be annotated as having 1,000,000,000 points even if it was not assigned to any user
__________________________________________________ ____________________________________________
Starting to work on the clients code will come later in time, first up is fixing the JSP jobs so that the lost stats may be joined with the current stats and fixing the system so that new libraries can be loaded. Only after this has been done will Andy start to look over the code of the client and of the servers, or so this is what I understand.
Where as all of us would like to see the stats recovered and joined with the current run, if it can't happen it is not going to happen; but it may be entirely possible to create a static page with the old run of stats on it but they will be static and we will continue to build on the new stats.
Many may not agree with this but what must come before the stats is protecting the science we have generated in our results. For TRI this is paramount, otherwise what is the point of continuing on? When there is a power loss at TRI there is finite amount of electricity that can be generated to run hardware on. While as it may be an inconvenience to you and I that the web server or stats server is off, the power is better used running the refrigeration units in the wet lab than the web sever. We would not want the samples to expire to do no refrigeration just so that we can get our stats fix for that day. The web server will come back up and the results will propagate to the stats server and everything will be updated, but the science has to come first.
You can at least expect new libraries in the future as well as a new client/backend. The stats as I said may end up just being a static reference page if they cannot be merged with the current JSP.
__________________________________________________ ___________________________________________
For the most part I agree with Rodzilla, but I think that the stats should stay lost. After all the alleged "cheating" (I'm not pointing any fingers here, so nobody get upset!)and other stuff thats happend recently, I think the stats should start off from new. This way ALL the teams are starting from an equal point and not being boosted by extremely high numbers from LOOOOONG since inactive nodes / members. Lets start with new stats and only active team members. This would be a MUCH more interesting game of chase.
There must be some middle point we could reach were everyone is happy, maybe if a new bug free / updated / same client - different number / whatever /insert here, client can be released fairly soon, we could restart the stats from that new client, and update the final stats for the previous (current) client to reflect the true numbers that the teams produced under that run.
If you look at the current team stats, you will see that there have been a LOT of position changes, and lots of non / low production teams are now much further down than they used to be. and some teams are right back at their position before the stats loss. These stats now reflect the true, current team production numbers.
Same teams, just new battles ahead.
Lagu
Once an AMDuser always an AMD user