Thanks for the news here first Matt. I have just received my first 2 WUs with the new quorum too.

May I suggest that it would be a good idea to put this on the front page news and RSS feed, for those that may not read the forums often.

Have you also considered an email announcement as well?

---

Sir Arthur C Clarke 1917-2008

May I suggest that it would be a good idea to put this on the front page news and RSS feed, for those that may not read the forums often.

Have you also considered an email announcement as well?

I think whoever takes such a close look at his wus to wonder about the change will also have a look at the message boards.

---

mic.

With flop-counting, most machines claim exactly the same credit. It'll only be an issue when a current BOINC client gets paired with an older BOINC client, and I think the lower score is usually thrown out, so it'll likely raise credit a tiny bit on average.

I agree entirely with your first remark, but not the last part: with only two results to compare, the lower claim determines the credit granted. And the pre-Flop-counting BOINC clients sometimes claim higher than current ones and sometimes lower, so the bias could go either way. AFAICT it will be very tiny, regardless; the proportion of hosts running old clients is quite small now.

---

PhonAcq
[…]
Why are you creating Hate and Discontent now?

How is it “creating Hate and Discontent” to politely request information about the basis for a project decision? Surveying this thread I see a mostly cordial and informative discussion, with only one person shouting accusations and making a fuss … and it isn’t PhoneAcq. (Just my 2¢ worth, take it or leave it.)

PhonAcq
[…]
Why are you creating Hate and Discontent now?

How is it “creating Hate and Discontent” to politely request information about the basis for a project decision? Surveying this thread I see a mostly cordial and informative discussion, with only one person shouting accusations and making a fuss … and it isn’t PhoneAcq. (Just my 2¢ worth, take it or leave it.)

PhonAcq has been pretty consistently questioning the management of the project. His theme seems to be that most decisions, like this one, are mostly random.

... and certainly, if SETI had a staff of 20 and a much larger budget, there would be a more formal review process, reports, meeting minutes, etc.

But SETI doesn't have a staff of twenty, it's closer to five or six, and the management is a lot less formal in a small project. It is an academic project, not a commercial project, so there are no shareholders who depend on this for their future financial security.

"Informal" is not a synonym for "sloppy."

It seems well thought out to me, and it seems that they're ready to measure the change and modify how things are done based on the result.

Edit: We also seem to forget that this is a litmus test: any signal we find is going to be verified and reobserved several times.

---

Redundancy level of one--- Does that mean that there are two results that agree? If so, and the probability that those two results are truly correct is 99 percent--- Does that mean that each unchecked result has a 90 percent chance of bring correct, and that three checks would imply a certainty of 99.9 percent? Of course it would seem that any false positive found in the results would easily be disproven by more results.

Hi! It is 99 for two results only if both false results are equall and validate.
This is unlikely: If results disagree new work is generated untill two
results are same.

What are real probabilities for false results entering database?

73

Iztok

His theme seems to be...

Perhaps you have more history behind the belief than I wish to go search for, but IMO, the questions initially raised and the follow-up seemed civil enough. As someone who recently fell victim to the prejudice of an untrue initial belief, I can tell you that it is quite frustrating to have only the negatives about you looked at and any positives either mostly or completely discarded...

FWIW and YMMV...

Brian

---

Anyone think about "Pending Credit", or tasks being incorrectly interpreted as "stale wus"?

We used to have the 3/2. It was changed to 4/3 for a few reasons; one being the slightly less chance of incorrect results, another being a reduction in the time a wu spends in the DB, but the MAJOR reason was so we participants could get our "precious" credit faster.

When 3 were sent out, it wasn't uncommon for one or two to have "download errors" (pre Boinc 4.19). There was no way the servers would know this until the wu timed out (ran past deadlines), so it would be two weeks (previous fixed deadline) until the server would issue replacements for those. Then the replacements would get two weeks to return it. Resulting in as much as a month before credit was granted, and sometimes even longer (this was even less frequent).

Now with varying deadlines it's possible to wait 108 or 162 days before credit is granted. This is if you get one with a 54 day deadline, and the second figure is if the first replacement isn't returned or is invalid.

Of course this isn't likely to happen frequently, but we should be aware of it's possibility. Also, be prepared for users "thinking" it's a stale wu, when in fact it's just the "normal" byproduct of the configuration.

tony

All good comments throughout the thread.

Questions can be asked in many different ways, ranging from the form of an "accusation" all the way through to polite and informal. For example, with the greatest of respect [you utter simpleton of a dork] some of the best barbed questions must be in public government sessions!

Such is the richness of innuendo in language and such is the scope for misunderstanding between translations and culture. Sometimes I'm left guessing when there's only the plain text to read! :-O


As for this thread:

Taking the assumption that the Science is sacrosanct and that the science database must be maintained 'correct' and verified 'correct', dropping the quorum count after so many years is a bit of a surprise.

Then again, it's all down to what is to be done with the Science. Berkeley certainly know that better than us!

My take is that reducing the minimum quorum to 2 is a good move to sacrifice very slightly the 'reliability' of the results in the science database to instead get the new ALFA data processed more quickly. The penalty is that Berkeley will need to do a bit more database bashing to weed out (or ignore) the few bad results that will creep in there. The whole lot will get reverified in any case with whatever reobservations runs are done to find whatever beacon or whatever.


To my mind, it's all very good science. It's also briliant juggling of very meagre Berkeley resources to maximise the science returns.

Certainly gets my vote of confidence (for whatever that's worth).

All good science and good fun!

Regards,
Martin

---

See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)

Special Thanks goes out @ Berkeley for a Science Well Done - Nice goin' ALL . . .

yes indeed thank you nobody .. I am pleased to be allowed to participate in this scientific experiment SETI@home/AstroPulse

Fellow readers, please accept my apologies for giving Pappa a chance to flame about a reasonable observation I made public. I received objective replies to my questions from CentralCommand as well as other readers, which we should all appreciate. In an objective organization, it is not a bad idea to look for ways to improve the process, even if it is not feasible at the moment. It's similar to the scientific process we all admire. And CentralCommand is pretty good about fielding critical observations.

Pappa, on the other hand, seems to have a raw nerve exposed. I don't know him, and have no desire to know him or his progeny. But I do apologize to the rest of you for giving him a chance to demogogue.

One observation:

Yes we will be able to anaylze more WU's in a given period of time at 3/2 than at 4/3, and that's a good thing.

However, under "ideal" conditions the efficiency from a power consumption POV is worse, since now 33% of the total results sent out will not be contributing to the science instead of 25%. Remember once a WU validates and the canonical is chosen every other result returned after that is fluff.

@PhoneAcq:

One thing to keep in mind about your concerns on redundancy is it takes 2 strongly similar results to validate, otherwise it will wait for another result to arrive. Theoretically, this should address most of the problems, since the odds would be host side effects would result in at best weakly similar results and cause validation to wait for the third result or force another result to be issued.

Since the app is open source (and the reason we have the great Coop and other apps in the first place) experience with the bug we saw with the early Coop 2x version makes a stong argument in favor of a minimum quorum of three to help mitigate the effect of problems like that, as well as the issue of allowing 4x clients and the BM scoring they use.

Alinator

This is a good opportunity for me to get on my "change control soapbox". Would someone give a good (technical, quantitative) reason for the change? Would someone give a good (technical, quantitative) analysis of the risk involved? It just seems like good project management to go to the effort to engineer the change (and good PR to publish the results).

And more recently:

Fellow readers, please accept my apologies for giving Pappa a chance to flame about a reasonable observation I made public. I received objective replies to my questions from CentralCommand as well as other readers, which we should all appreciate...

[RANT ALERT!]


Sorry PhonAcq, but to my understanding of English and innuendo, that is a complete "arse about tit" supercilious non-apology.

So, do you indeed know better than everyone else here? From the way you give your opinions in your questions, you very definitely give a very strong "high 'n' all-mighty" omnipotent view that you already know the answer.

So, what in reality do you know?...


Ignoring the condescending approach with your questioning, the points of your questions are fine. The manner of your questions however appear very patronising and verge on being insulting.

Or is English not your first language?

(Or I could phrase it that "your communication skills are not good are they?" With whatever respect! (Can you see the multiple puns in there, and in this?...))


I guess people are human and need to learn further, including you.

Happy crunchin',
Martin

---

See new freedom: Mageia Linux
Take a look for yourself: Linux Format
The Future is what We all make IT (GPLv3)

However, under "ideal" conditions the efficiency from a power consumption POV is worse, since now 33% of the total results sent out will not be contributing to the science instead of 25%. Remember once a WU validates and the canonical is chosen every other result returned after that is fluff.

Hummmm... I think a more appropriate point of view is that the power efficiency is considerably improved. In the most common case, each WU will consume the system power to be computed three times instead of four. The extras beyond the first generate confidence. Confidence beyond that needed for the scientific result should not be weighed equally in a power efficiency calculation.

Obviously I am assuming the project folks are making a good choice on the value (or lack of value) of quorum-driven confidence.

Einstein has been sending out two and requiring two for some time. That combined with a very non-random quorum partner assignment gives a much more bursty pattern of validation and credit assignment, but most of the regulars posting there seem to have found their peace with it. It also is a considerable additional power efficiency and total output gain over 3/2.

---

Fellow readers, please accept my apologies for giving Pappa a chance to flame about a reasonable observation I made public. I received objective replies to my questions from CentralCommand as well as other readers, which we should all appreciate. In an objective organization, it is not a bad idea to look for ways to improve the process, even if it is not feasible at the moment. It's similar to the scientific process we all admire. And CentralCommand is pretty good about fielding critical observations.

Pappa, on the other hand, seems to have a raw nerve exposed. I don't know him, and have no desire to know him or his progeny. But I do apologize to the rest of you for giving him a chance to demogogue.

Setting aside the fact that this is less about the apology and more about trying to discredit Pappa.....

It isn't so much about asking reasonable questions than a question of style.

It is possible to ask about the effects of a change without calling the the decision process into question.

Change control is good. The fact that SETI doesn't have a formal Change Control Board doesn't mean they aren't doing it.

---

Einstein has been sending out two and requiring two for some time. That combined with a very non-random quorum partner assignment gives a much more bursty pattern of validation and credit assignment, but most of the regulars posting there seem to have found their peace with it. It also is a considerable additional power efficiency and total output gain over 3/2.

I didn't know this about Einstein ... sending out two and requiring two ...would this also work better for SETI@home ?

Folks, I was trying to be sincere. Sorry you took it differently. (Except my crack about Pappa, who seems to me to have started this e-tussle.)

@Ned: I'm not sure what it is in what I originally posted "calls the decision process into question", at least in a negative way. I was asking two objective questions ( 1. what is the anticipated (quantitative) benefit, and 2. what are the anticipated side-effects and possible omissions; again with a quantitative estimate.). I suppose I asked implicitly a third question (why not make the project leaders' reasoning visible- quantitively with a minimum of hand waving). So I think your conclusion about my intentions and the like is wrong. And the good news is there was some good discussion and illumination of the points as a result of my question.

@Martin: I wonder who (you or me) comes across as more supercilious. But it doesn't matter to me if you think I'm the best.

A better board would be filled with objective analysis of the issues, rather than amateur psycho-analysis, emotional demagoguery, and hand-waving argumentation. Let's all just try to be better engineers and scientists.

Surely the amount of crunching, and hence electical power, that is wasted, is relative to the percentage of times the validation process finds a strong match on the first two tasks that it compares. If this over 50% of the time then one could argue that in most cases the third, forth tasks are wasted resorces.

It might be better to delay the issuing of the third replication of the WU for 72hrs. I get this figure from my tasks shown in my account at the moment;
total 211
18 crunching or waiting to crunch
33 pending of which 24 (72%) were issued in last 72 hrs.

And since the change (between 10:40 and 11:30 yesteday) to 3:2 only 4 out of 24 tasks reported are on the pending list, and most of tasks granted credit have only two tasks returned.

---

I appreciate everybody's concerns and questions at every level. I've been around the block, and met many fairly unreasonable humans, none of whom are represented here in this thread, so any of this (very minor) contention was unnecessary and hopefully we can just drop it.

- Matt

---

-- BOINC/SETI@home network/web/science/development person
-- "Any idiot can have a good idea. What is hard is to do it." - Jeanne-Claude