Closed source is fine by me, howether to take away a measure not only of team V team but of your own personal endeavours would be foolish IMHO.

AT UD we had a pure and simple measure, WU`s done, yes some were longer than others but it evened out over time.

A measure of some kind is beneficial to the project and to most users.

---

Not all Czech`s bounce but I`d like to try with Barbar ;-)

Make no mistake This IS the TEDDIES TEAM.

So Morphy, will you join me in asking for a BOINC free Rosetta or if not for a closed source BOINC with heavy encryption?

YES

Or better still: wil you join me in calling for credit-free DC Projects? Hey , it is the science that matters , not the credits :)

NO :)

---

Teddies....

So Morphy, will you join me in asking for a BOINC free Rosetta or if not for a closed source BOINC with heavy encryption?

YES

Or better still: wil you join me in calling for credit-free DC Projects? Hey , it is the science that matters , not the credits :)

NO :)

At least you are intelectualy honest. :) :) :)

I understand this, but what is to be gained by using the averaged credit claims of many computers versus the average time claims? The former can be 'manipulated', the latter seems to be more of a constant. Time can also be judged by the scientists. . since the simulation crunch time isn't affected by client version, bogus claims of time are more easily contested than client benchmarks.

The "claimed credit" formula is (whetstone+dhrystone) X cpu time (reported)/ 1728000. Almost all optimized Boinc clients change the benchmark values, but Trux 5.3.12tx36 and 40 change the reported "cpu time" to adjust the claimed credit. Trux was nice enough to let the user decide whether or not to turn the "calibration" on for each project. Unfortunately, either knowingly or unknowingly users leave the calibration "ON" for projects that don't have an optimized app available.

Does Rosetta provide an "fpops estimate" with each WU? I'd assume they do since it's used by in determining the DCF and thereby the "estimate to completion". Is it possible to have Rosetta provide an "Fpops Estimate/Model" in addition to the "Fpops est/WU" (if it exists)?

tony

Hey Carl.h! long time!..........credits credits credits.........I'll keep running however things turn out.

I used trux's Boinc Client when I was using the optimized SCIENCE application for Einstein, because without I was bringing the Claimed Credits of the quorum DOWN from the standard.......Use the standard client now cause I'm using standard apps.....just the path of least resistance for me......it's not really a big deal to me.......credits are kinda of blahh

What I do miss, and what actually held real value for me in past projects (FAD, and even Grid.org, was total CPU time......and results returned...

those values made me feel good........when I could say I've done x number of workunits that is equivalent to x years of 24/7 computer time.

back with FAD we all did about 11,000 years of crunching......grid is at the equivalent of 475,000 YEARS of crunching.

that has the meaning to me.

wish BOINC tracked those values.

The fact I like Rugby Union does not mean Rugby league (nancy game) is illegal.

Very true but a try is worth the same amount of points regardless of which team scores it (in the same code). That doesn't mean that each player is equally capable of scoring a try, it doesn't preclude that fact that some players have natural abilities and natural advantages over other players, but a try is a try is a try.

And we have common ground on the choice of codes :)

---

Just to say something (no aggressiveness, no flame, no war) :

What is interesting in the credits (in my point of view), is to give an idea of the work you have done. I don't care to be at the second or the 2 million place. But I'm interested to know after one year how much work I've done. Credits don't give you really an idea of what you have done because of the different CPU, apps, clients, ... .
Why not use two indicators? One saying how many WUs have been crunched and a second with "credits", based on CPU time or whatever.
In Seti classic 1 WU = 1 'point', it was easy (but all WUs had the same lenght). But you knew exactly what you have done.

Hi Matt great to see you still about !

Yes the amount of crunching time is missed and possibly means more than credits, well it did when it was based on # WU`s.

The driving force for a lot of us is forever upward, try to do that bit more than yesterday or a teamy, or even an opposing team ! Without some sort of accreditation system there would be no competition and no drive.

There has to be a measuring stick and I guess it can`t be open source, playing with it is obviously causing some concern, hopefully though those days are numbered.

Matt do pop in to the Teddies for a chat sometime, I miss the good old debates though possibly we went through most and agreed very little but it was good exercise for the grey matter !

Keep well mate !

---

Not all Czech`s bounce but I`d like to try with Barbar ;-)

Make no mistake This IS the TEDDIES TEAM.

Trog Dog so a try against the Barbarians or the All Blacks is the same as a try against the White Horse Pub ?

Fact : A new accreditation method will be implemented very shortly.

There is no need to carry on and on over optimised clients. No one is winning and the forum/project is losing.

---

Not all Czech`s bounce but I`d like to try with Barbar ;-)

Make no mistake This IS the TEDDIES TEAM.

Just to say something (no aggressiveness, no flame, no war) :

What is interesting in the credits (in my point of view), is to give an idea of the work you have done. I don't care to be at the second or the 2 million place. But I'm interested to know after one year how much work I've done. Credits don't give you really an idea of what you have done because of the different CPU, apps, clients, ... .
Why not use two indicators? One saying how many WUs have been crunched and a second with "credits", based on CPU time or whatever.
In Seti classic 1 WU = 1 'point', it was easy (but all WUs had the same lenght). But you knew exactly what you have done.

CPU Time and Models Created are great standards in which to base a Credit system ( They do measure real work produced and time contributed) they are like a good cake...tasty , substantial vs the curent "credits based on benchmarks" systems which is fluff. I think we can all agree on that.

BTW Tony : the developers do have estimates of Fpops for each computer host and they can update them, should the need arise. The data is there and can be used to generate a work produced based credit system.

BUT again, let's not delude ourselves the new system will solve the issues that have been raised. As long as it is set in and used in an open source platform/application, ANY system will be vulnerable to tampering. So not only is a new credit system is needed (and thanks be to G-d it will be comming soon) a secure application in which to run that credit system is needed and alas ,as it stands now BOINC, is not.

So lets wait and see what our developers present as the new credit system. Meanwhile, lets crunch till we drop :P

I agree with the consensus of the last few messages...........figuring out a new credit method is good.........but the root of the problem is that things are open source.........no matter the system.....if it's open, and available, someone will adjust it.......even if "just for kicks".....which will devalue the meaning of the numbers......if it's not standard and fixed.....it's not much of an indicator of anything because you can't be sure the values are valid.

I believe Einstein has made a move in the right direction, even if the system isn't perfected yet.......the credits are granted on the projects server side, and have nothing to do with BOINC's figures..........no matter which BOINC client one uses......the same credit is applied.......which I think if I understand correctly is what Rosetta is working on too......

removing the crediting system from the open domain where "we" can't fool with it is the only way..........projects doing it themselves seems like the first step........but that would require some serious cooperation between different projects to get the values to be in alignment.........but I don't think is asking too much.

Trog Dog so a try against the Barbarians or the All Blacks is the same as a try against the White Horse Pub ?

Fact : A new accreditation method will be implemented very shortly.

There is no need to carry on and on over optimised clients. No one is winning and the forum/project is losing.

Dear Lord!!! You have cools names in your team :)

I agree with the consensus of the last few messages...........figuring out a new credit method is good.........but the root of the problem is that things are open source.........no matter the system.....if it's open, and available, someone will adjust it.......even if "just for kicks".....which will devalue the meaning of the numbers......if it's not standard and fixed.....it's not much of an indicator of anything because you can't be sure the values are valid.

I believe Einstein has made a move in the right direction, even if the system isn't perfected yet.......the credits are granted on the projects server side, and have nothing to do with BOINC's figures..........no matter which BOINC client one uses......the same credit is applied.......which I think if I understand correctly is what Rosetta is working on too......

removing the crediting system from the open domain where "we" can't fool with it is the only way..........projects doing it themselves seems like the first step........but that would require some serious cooperation between different projects to get the values to be in alignment.........but I don't think is asking too much.

Want an easy first step to reduce the tinkering with the credit system: Disable the capacity of a participant to reset/recalculate his benchamrks at will. Follow that by making the resetting of benchmarks a random action that originates in the project and do not allow the benchmarks to remain active for more than 24 hours cpu time.

As long as you use anything thats tied to any benchmark whereever it comes from your going to have issues.
Benchmarks, even the best of them and the BOINC benchmark isn't one of the best, only serve to show theoritical capabilities.
If you want an honest scoring system, you have to go to something that works off of the actual work done. It also has to be "fair" to Intel,AMD and Macs.
It doesn't matter what values are given to a certain amount of work as long as it's given to all..1 point or a 1000 isn't an issue as relates to Rosetta. It may be an issue as relates to other BOINC projects but quite honestly I don't care about them and neither should Baker Labs.
I can tell you this for a fact: Go to a work based credit system that eliminates the current loopholes and I can bring hundreds of new people to this project that have stayed away JUST because of the fact that there are so many ways to manipulate the points. I'm not talking someone who accords Rosetta 15% of their computer time every 3 days but people who will bring multiple machines that will run 24/7 on this project.
These people would more than make up for any that left the project due to any issues between Rosetta and BOINC.
Maybe the only answer to this is to take Rosetta off of BOINC altogether or we will never hear the end of the complaints of Rosetta's scoring effecting other BOINC projects.
I really do not understand why this is so hard for some to understand. Keep it simple. Something as easy as one point per decoy returned and the speed of the machines will sort out the rest all by themselves.
If one point doesn't keep the total "values" where people are happy, make it two, or 22..It is of no importance as long as the amounts awarded are in relation to the work done.
Some mentioned doing away with the credit system alltogether.
If that was done. Rosetta would lose 50% of the current computational power that is directed to it.
People like to compete and if you tell them that your removing the competition they will go somewhere else to find it.
Like it or not that is the brutal truth.
The points and the competitions serve to add a little "spice" to what would be a very boring exercise without them.
They also serve to bring new people into Rosetta.
Bottom line: Come up with a simple system that awards points based on work done and you are home free.
I feel like I've said this so many times that either I'm the stupidest person in history or I don't know how to say it to get across my meaning.
If anyone doesn't understand what I am saying, please say so. Otherwise tell me what is wrong with my thinking. If not, why all the discussion, just implement it and lets get on with the next issue.
Thanks for your time,
A very frustrated Movieman

---

Yes I think I understand perfectly what you're saying. I must say that this looks to be the only viable option left. I think that Ethan was trending in this direction as well.

I also agree that some type of credit system needs to remain in place since competition does draw in many.

---

I feel like I've said this so many times that either I'm the stupidest person in history or I don't know how to say it to get across my meaning.
If anyone doesn't understand what I am saying, please say so. Otherwise tell me what is wrong with my thinking. If not, why all the discussion, just implement it and lets get on with the next issue.
Thanks for your time,
A very frustrated Movieman

David : you are not stupid...you are the most dedicated Rosetreer I know. I for one have to thank you for having hooked me into Rosetta.

Te he te he (Insert blushing Smilie here) I was the one that proposed doing away with the credits. I was playing the devil advocate.., I just wanted to show that credits fluffy are they are do have a role in the DC Community. :)

I agree with you ...a credit systems that measures work produced, simple and secure is the way to go.

Thanks for caring as you do for the well being of the project.

Yes I think I understand perfectly what you're saying. I must say that this looks to be the only viable option left. I think that Ethan was trending in this direction as well.

I also agree that some type of credit system needs to remain in place since competition does draw in many.

I think there is a consensus that a simple safe credit system based on work produced is the best way. I think that it is also agreed that the creidts spice up the projects, although they should not be the most important part of the projects.

I feel like I've said this so many times that either I'm the stupidest person in history or I don't know how to say it to get across my meaning.
If anyone doesn't understand what I am saying, please say so. Otherwise tell me what is wrong with my thinking. If not, why all the discussion, just implement it and lets get on with the next issue.
Thanks for your time,
A very frustrated Movieman

David : you are not stupid...you are the most dedicated Rosetreer I know. I for one have to thank you for having hooked me into Rosetta.

Te he te he (Insert blushing Smilie here) I was the one that proposed doing away with the credits. I was playing the devil advocate.., I just wanted to show that credits fluffy are they are do have a role in the DC Community. :)

I agree with you ...a credit systems that measures work produced, simple and secure is the way to go.

Thanks for caring as you do for the well being of the project.

Who cares about the "cheating" of an optimized client. First of all, a fix has been promised after CASP is finished (last expiry is 7-AUG-06), and right now that is (or should be) the primary focus. Secondly, it has no bearing on the science being done (claim 10 credits or 1000 credits, valid results are what matters).

My only comment relative to credit is that the number of decoys (as some have suggested) is not a sufficient qualifier. It also needs to take into account the number of proteins since the larger the number of proteins the longer it usually takes per decoy.

Let's just wait until after CASP and see what the programmers have come up with for an equitable, fudge-proof (hopefully) credit system.

Can't we all just get along and help the science ;-)?

---

!!Stupidity should be PAINFUL!!

Boinc reform should be presented at the Boinc development forums.

Let's get back to the Rosetta credit system - which is something the Rosetta team can do on their own.

One of the reasons for the recommendation for determining a credit/model for each major WU on Ralph (or Rosetta Labs) was to get around the problem of fixed credits per model; where the models from different WUs take vastly different amounts of time. My low amount was 11 models a day in this last week. My high was 53 models a day in this last week. I don't look at the values that often, but I've seen numbers close to 180 models a day in the past - and we've worked on much faster WUs than that in the past.

While a fixed number of points per model may balance out over time, the temptation for the credit freaks would be to benchmark each of the current WUs - and then dump all the 11 model a day WUs, and only work on the 53 model/day WU (perhaps the generous would crunch some of the 30/day or higher WUs). By giving each major WU a score for the models based on how long that WU takes to create a model, we don't have to worry about people cherry picking which WU to crunch.

Back to Ethan's question - what do we consider an adequate number of machines or cross section of machines to be used to generate a Rosetta only score for each Rosetta WU?

---

Who cares about the "cheating" of an optimized client. First of all, a fix has been promised after CASP is finished (last expiry is 7-AUG-06), and right now that is (or should be) the primary focus. Secondly, it has no bearing on the science being done (claim 10 credits or 1000 credits, valid results are what matters).

My only comment relative to credit is that the number of decoys (as some have suggested) is not a sufficient qualifier. It also needs to take into account the number of proteins since the larger the number of proteins the longer it usually takes per decoy.

Let's just wait until after CASP and see what the programmers have come up with for an equitable, fudge-proof (hopefully) credit system.

Can't we all just get along and help the science ;-)?

Well... tell that to the people that came in this MB to tag as cheats aall people using optimized clients. The brouhaha was not the result of an immaculate conception: people came in and cried cheat and tagged people as cheats.... and people responded to them/

On you comment re the decoys: I agree with you and I am waiting to see how the issue is dealt with by the developers.

On your final comment: I agree with you.