![]() |
![]() |
#1 |
Sep 2003
Borg HQ, Delta Quadrant
2×33×13 Posts |
![]()
We haven't heard anything in a while, so I'm just wondering, is the v5 project continuing, is it dead, or what?
|
![]() |
![]() |
![]() |
#2 |
Jul 2004
Nowhere
80910 Posts |
![]()
It will be here eventually
----BUMP---- |
![]() |
![]() |
![]() |
#3 |
Sep 2002
23·37 Posts |
![]()
bump
|
![]() |
![]() |
![]() |
#4 |
P90 years forever!
Aug 2002
Yeehaw, FL
2×3×7×199 Posts |
![]()
Scott and I have been updating the client-server protocols for the last month. Scott has begun writing server code, I've started work on putting the communication code in its own thread in preparation for v5 changes.
Coding and testing will take quite a while and there are still some issues to resolve especially in the area of converting v4 clients and accounts. Also stats reporting is not fleshed out. When the time comes, be prepared for transition pain. V5 has users with users belonging to teams. V4 userids represent either a team or a user. So what do we do with a v4 team's CPU credit. We can't give that credit to just one user of the team and we don't know how to split it up among the users in the team. Thus, we cannot accurately rank the CPU credit produced by all users. Does this mean stats need to be reset to zero? Last fiddled with by Prime95 on 2005-10-03 at 00:07 |
![]() |
![]() |
![]() |
#5 | |
Aug 2002
21EE16 Posts |
![]() Quote:
![]() |
|
![]() |
![]() |
![]() |
#6 |
P90 years forever!
Aug 2002
Yeehaw, FL
2×3×7×199 Posts |
![]()
I'm leaning toward some sort of plan where v4 stats carry forward and folks are more or less on the honor system for dividing up team stats into individual user stats and on the honor system for linking v5 and v4 accounts.
However, the stats page would not place much emphasis on the lifetime contribution rankings. Instead, maybe year-to-date or v5-startup-to-date contributions would be the primary focus. The advantage of highlighting year-to-date rankings is that it encourages current contributions. No final decisions have been made. Comments are welcome. |
![]() |
![]() |
![]() |
#7 |
Jun 2005
USA, IL
193 Posts |
![]()
I would like for the lifetime contributions to continue. I would rather only see any YTD or V5-STD contributions as more of an addendum or additional column to sort by.
![]() |
![]() |
![]() |
![]() |
#8 | ||
Aug 2002
Dawn of the Dead
111010112 Posts |
![]()
Pain is inevitable. It is prudent to plan this now, in order to minimize the level of suffering.
There is no need to rip apart the guts of old teams. Splitting up the credit of old users is a hazard you must avoid at all costs. I recommend you leave things as they are and carry forward, as Step II, to v5. So, TeamX is still composed of its members user_A, user_B etc.; the team simply crunches on without anyone being aware of the change. You are most certainly thinking about improving the stats to encourage more of the competitive type DC teams to field an effort. All other DC projects are structured around the user's individual stats. A user account can be comprised of one or many machines, with no distinction between the individual units. There exists a mechanism for joining an individual account to a team. The stats system is capable of reporting both data: The individual tally, whether on a team or not, and the team rankings. The difference in GIMPS is in the individual machine ID. One could deploy many machines under a common ID. The team function allowed super-users to track individual machines and competitive multi user teams form the same way. At this time we're in Step III. In the end, the difference is in the end user. A general purpose team that discovers GIMPS due to the newly viable stats won't care that individual tracking is lost. Such a team exists for one purpose: The thrill of the chase. Yet, in the end it matters little, especially to you. This is because large competitive teams host their own statistics engines. For example, Team Prime Rib has an elaborate system where we use a naming format to assign individual machines to a single user account. See my status on the TPR system for an example of what can be done. Even if you abolished individual ID in favor of the concept of general user, we would still continue as before - because our system, not the project's, performs this function. Even if the project forced me to reassign all my machines into a general account, it would make no difference whatsoever to my long term stats history, as our statsmeister would simply add a redirect adding my new general user to my TPR account. A side note for v5: make the general stats report simple, such that team stats engines can easily parse the data. Another side note: Do not worry about much of stats details on your end. Big teams comprise hundreds or thousands of users, amongst these IT pros, programmers and other keen and capable enthusiaists who make stats, write helper apps, break legs etc. Quote:
You don't even want to dream of going there. This is the stuff that destroys teams and harms projects. See Genome@Home, Ars TPS vs. [H]ardOCP for an example of what a stats fiasco can do to a project. You might recall this as you began posting in our thread at Ars around that time - TPR grew out of the remains of that debacle. Quote:
Last fiddled with by PageFault on 2005-10-07 at 02:29 |
||
![]() |
![]() |
![]() |
#9 |
Jul 2004
Nowhere
809 Posts |
![]()
what if old accounts carried forward with the same credit so say tpr would have what they have now then users would have to make there own accounts and join back to tpr that way the teams keep credit and then there is a fair way to treat it. Team passwords are the same as they account pass that came from v4 so they can still be controled.
|
![]() |
![]() |
![]() |
#10 |
P90 years forever!
Aug 2002
Yeehaw, FL
2·3·7·199 Posts |
![]()
It is easy to have teams retain their credit. What is hard is users. If I don't let tpr tell me which users produced their credit then the tpr users are at a disadvantage on the "Rank all users stats page".
|
![]() |
![]() |
![]() |
#11 |
Jul 2004
Nowhere
809 Posts |
![]()
no i mean make all v4 users a team with a new pass that is emailed to the user with the email address listed then all users are not teams are set to 0
so it goes v4 user credit --> version 5 team credit v4 pass changed --> emailed to listed address in account v5 users --> remake there account so everyone is at 0 |
![]() |
![]() |
![]() |
Thread Tools | |
![]() |
||||
Thread | Thread Starter | Forum | Replies | Last Post |
Mersenne.ca Status Report | Gordon | mersenne.ca | 168 | 2018-06-17 22:48 |
Exponent status report | Dubslow | PrimeNet | 2 | 2015-10-05 05:21 |
Mersenne.ca Status Report | Gordon | mersenne.ca | 1 | 2015-09-22 10:53 |
"Classic" (colourful) Status report | davieddy | PrimeNet | 6 | 2009-10-04 08:33 |
Delayed status report (split from main reservation thread) | rogue | Sierpinski/Riesel Base 5 | 8 | 2006-03-04 13:59 |