...
I guess, that within an early stage of a new project, all sides (project-team and crunchers) have a lot to learn, making the project really work fine.

But after this phase, there is a big change: The questions, stated from crunchers, keep to be the same questions as in the month before
...

This is right. And there are two other factors that re-inforce this.

In the early days of a project, the project team are motivated also by the fear that they will not get enough crunchers to achieve the mimimum level for a viable project. I think I read somewhere that currenlty Rosetta is around 10% - 15% of their minimum viable level of cpu power. So good PR is partly a matter of strong self interests at that stage. Later on, more cpus may be good, but is no longer a matter of project viability.

Secondly, in thae later stages of a project, the scientists have interesting results to publish. If the PR has been successfull in recruiting far more donors than the minimum, the scientists will have results coming out of their ears. In terms of benefit to science, these results are no use unless they are properly written up and published. In terms of benefit to the individual scientist's career, they are no good unless published before equivalent results from some other team (maybe DC or maybe mainframe based). So there is a strong attractor away from the message boards.

The four factors, more posts to keep up with, growing boredom with repeated newbie questions, reducing motivation for communication with users, and increasing motivation for communication with other scientists all reinforce one another.

You can't beat it. And in classic internet style, what you can't beat you work around. The work around for the reduced-communication syndrome is, in Yeti's words

to have some experienced volunteers, that keep an eye on the forums; yeah, let them be responsible for a part of the forum. If they can't answer a question, they direct a project-member to postings, that should be answered by them.

In a list of points on how to _continue_ to have the best BOINC project, promoting volunteers who show themselves to be capable and motivated to positions like forum moderator, etc would be an important factor. There still need to be some postings by the scientists, but they spread further if supported by suitable volunteers.

This explains both why Predictor once was best in communication, and also why it no longer is -- if they'd tried/succeeded to get keen volunteers the departure of one individual would not have had such an impact. In my opinion.

River~~

... we are both dinosauriers: My first contact with computers has been in 1977

dinosaurs?

young upstarts more like, to those of us who started in 1966 and can read 7-bit ASCII straight off paper tape, who memorised the Hollerith card punch codes...

and most importantly who remember putting in a card job and having to wait for the next "batch" to run. Who remember how long debugging took when you only got a few runs per day...

but we have been modern, we didn't need to punch, we only needed to make a sign on the numbered card ;-)

early mammal... maybe a Yeti ?


;-)

can read 7-bit ASCII straight off paper tape, who memorised the Hollerith card punch codes...

and most importantly who remember putting in a card job and having to wait for the next "batch" to run. Who remember how long debugging took when you only got a few runs per day...

early mammal... maybe a Yeti ?

PDP-8, paper tape... Xerox Sigma IX 'mainframe', 80-column cards and overnight batch runs... 96 column cards on IBM System III... HP KSR 33 keyboards... thrilled when 300 baud modems came out... "real programmers use assembler"... He may be a Yeti, but I still qualify as reptilian. :-(

---

Binary light panels, typing in "bootstrap" code to pull the OS off tape ...

Floor sorts ...

---

... Binary light panels ...

programming the front panel as chaser lights at Christmas,

programming them to show the wall time
and being able to read the time even tho it was in binary

Machine code programing on an Elliott 803 (4K ferrite core store) where the back-up medium was punched paper tape!

---

Uuh, this seems to be a collection of oldtimers ;-)

Maybe we should open a club ;-)

---

Supporting BOINC, a great concept !

Rotating magnetic drums for memory, vacuum tube 'flip flops'(2 on a card even!!) and troubleshooting using machine language, logic flow diagrams and Boolean equations.....ahhh the good old days.

---

The four factors, more posts to keep up with, growing boredom with repeated newbie questions, reducing motivation for communication with users, and increasing motivation for communication with other scientists all reinforce one another.

Getting back to the topic ...

River~~

I do think you hit the problem on the head, this is the best summation I have seen. But, I will have to admit that the first attitude recorded by a project person is along the lines of "we have important things to do ..."

If maintaining participant interest WHEN YOU DO NOT HAVE ENOUGH PARTICIPANTS, is not important, if not supremely important, then what is? WIth out the participants, no computers, no computers, no research, no research, no papers.

Yes, writing papers is important, doing research is important, and there are a lot of other things that are important. Here comes the "BUT", every paper talking about volunteer DC that I know of stresses the importance of engaging the participant community.

Now, it is likely that I am reading too much into the statement. Yet, ...

Predictor@Home is "bleeding" participants. How badly is hard to tell. There are still attracting new so it may be "who cares" for them. The brutal truth is that individually a participant does not matter. But, the attitude of the project does. Will predictor get a lot of my CPU time? Probably not. I will likely continue to support them at a very low level for a long time (currently 1%), but, will they even notice? Not likely ...

UCB's interaction with the participant base is, in my opinion, appalling. And, this may bite them in the rear. At the moment they have more participants than they can support almost. But, the enhanced SETI@Home application, AstroPulse, multi-beam, and other telescopes are all coming on line ... what then?

Worse, though the pool of participants continues to grow, and I, and others add machines, there are also more and more BOINC projects (and non-BOINC too) that are comepeting for this pool of participants.

Volunteers can only do so much. Bill and Tony as help desk people is simply marvelous. I try to do my part (though recently ... :( ) too. But, I cannot make up for project apathy, neither can BIll and Tony.

I guess I just need to put my "secret agenda" aside and take what comes ...

---

Another example of appalling behavior to volunteers is United Devices (Grid.org), a non Boinc. They startred out with cancer crunching which attracted large numbers (in 2001). Recently however they seem to have run into problems and are just sending out repeated old WU's. Members keep on asking for info but there is very little forthcoming. You are lucky if a company member comes on once per week, and only to give platitudes. I was a Mod there for over 4 years and we were also kept well in the dark.
One cause could be, I suppose, they are a commercial company and cannot/willnot discuss sensitive info that they believe will impact on their business. They also are bleeding (knowlegable) members at a high rate, those who just load and forget are probably not affected.

---

Is it okay for me to suggest that there could be a parallel manifesto about how to be the best BOINC particpant?

Especially on the forum. People seem to just jump in and create a new thread without looking at what's been written before. With "no new work" for instance, there's at least 3 threads with similar titles.

• Cannot get any work from rosetta ??
  
• No work from project
  
• No work from project.
  

The FAQ Hugo has been working on might help a little, though some people don't read anything unless it's an answer to their own question in their own thread.

A belated response to Webmaster Yoda... You enunciate one of my pet peeves, too. Another pet peeve is that a msg thread title is, let's say "No New Work" and half way through the thread it transcends into an unrelated discussion on the effects of disk space preferences or even into a discussion on cheating and the awarding of credits through bad benchmarking or optimized clients/apps, etc. It's hard for a participant to peruse the threads to look for a problem similar to their own when the answer might be buried in a thread whose title is unrelated to your problem.

I don't think there is a fix (besides education) to thread hijacking. My pet peeves are threads that are in the wrong place entirely ("help me I can't get work" in Science, or "team recruiting" in NC...) or threads with worthless titles ("Help!").

As a moderator I can _ask_ people to move a discussion elsewhere, but if there is a way for _me_ to move an entire thread or rename it (short of creating a new one and moving all the messages to it, one by one...) I haven't found it. Does any other mod or staff member know of a way?

I'm looking forward to the "How to be the best BOINC participant" output though. That's one I'll gladly "sticky"!

---

...
If maintaining participant interest WHEN YOU DO NOT HAVE ENOUGH PARTICIPANTS, is not important, if not supremely important, then what is? WIth out the participants, no computers, no computers, no research, no research, no papers.

This falls into two cases, both of which occur in thelife of a project

a) when you do not have enough participants nor enough results for even the short term,

b) when you do not have enough participants for the long term but you *do* have enough results for the short & medium term

In case (a) even the most donor-hostile scientist will be scrabbling to the forums to drum up support -- they aren't stupid.

In case (b) you run up against the ancient human dilemma of how to allocate priorities between the short and long term. In addition there is the fear that by spending less time on the results, other teams may get in first with results that you already had on hard disk, getting through the backlog of the existing results can be a powerful motivator.

There are two ways to figure out an answer to that dilemma.

Treat it as an exercise in games theory, and you spend just enough time on the forum to collect just enough data to fill the rest of your time. That may or my not be what the users want -- but it is a feedback loop, after a while you will be left putting in just enough effort to keep just enough users who are content with that amount of feedback. Say you believe in supporting the donor community, but actually do only as much as you need to. Machiavelli could figure it out.

The other, more friendly, factor is to notice that the greatest scientists have always loved communicating their field with others. Think of Einstein (who once took time out from his day job to help a neighbour's six year old with her sums, unknown to the girl's mother. When the mother finds out she is horrified and goes round to apologise. "Do not scold the child", Einstein said, "I learnt far more than I taught". Feynmann and Hawking are other 20th century physicists who were / are great at physics and great at communicating it too.

The scientists who feel like that will tend to stick around on project forums longer than the others. Their project will continue to communicate with donors far longer and far more thoroughly than others, not because it "pays" to do so, but simply because the scientists *enjoy* sharing their work with lay people.

These turn out to include the very greatest scientists too - Machiavelli would never figure that one out.

So what I am on the look-out for are *not* scientists who communicate with participants because it makes sense in a calculating way -- they will evaporate into their publications when the data arrives. What I am on the look out for is a project with scientists who seem actually to enjoy the engagement.

How to have the best BOINC project?

Number ONE: Pick scientists who see communication with participants as an enjoyable opportunity rather than a duty to appease the donors of cpu time.

It echoes the traditional wisdom of placing research in a teaching university rather than in a student-free research institute.

---

I don't think there is a fix (besides education) to thread hijacking. My pet peeves are threads that are in the wrong place entirely ("help me I can't get work" in Science, or "team recruiting" in NC...) or threads with worthless titles ("Help!"). ...(edit: remainder deleted)...

I agree very much with every point you make here. Ahem, however, as much as I respect you, Paul, and others, I must point out that several of you digressed in this thread to reminisce about the "good ol' days" of 80-col punched cards, floor sorting, RPG, overnight batch jobs, etc. I am a retired computer programmer and systems analyst(IBM mainframe, COBOL) for Rockwell International, but I have no interest in the "good ol' days". I've moved on to DC! The "good ol' days" is a topic for the Cafe perhaps. Sorry. :)

I must point out that several of you digressed in this thread

Mea culpa... new thread started in cafe, please see it there as I have a question for you... :-)

---

Is it okay for me to suggest that there could be a parallel manifesto about how to be the best BOINC particpant?

Please... PLEASE! Today I really need one like that. :-)

(And I'm not just referring to Rosetta issues/participants, but also a couple of other boards...)

---

Chin up Bill, I was a Mod at UD for 3 years.........

---

Is it okay for me to suggest that there could be a parallel manifesto about how to be the best BOINC particpant?

Please... PLEASE! Today I really need one like that. :-)

(And I'm not just referring to Rosetta issues/participants, but also a couple of other boards...)

Well, the way to reduce the heat in the Rosetta kitchen is to get rid of the source: bad WUs and hastily implemented science app. People are frustrated. But you are right that, having explained the problem on the boards, why aren't participants just accepting that what's done is done, and crying over and over is just juvenile. I'm also interested in what the project devs think comprises a good DC participant. I would like to see that manifesto.

Well, the way to reduce the heat in the Rosetta kitchen is to get rid of the source: bad WUs and hastily implemented science app.

NO argument on those!

People are frustrated. But you are right that, having explained the problem on the boards, why aren't participants just accepting that what's done is done, and crying over and over is just juvenile. I'm also interested in what the project devs think comprises a good DC participant. I would like to see that manifesto.

I don't know what the devs would think - MY list would start with "Read the front page, the technical news, and then all the recent threads, and at least do a search for your problem, before posting. Your problem may not be unique, everybody may be having it."

Next would be "Look at the thread titles on the first page of threads. If one of them is exactly what you're about to put, or even close, consider adding to _that_ thread instead of starting a new one."

And there's always "Are you even in the right place? Not only the right Forum, but the right Project?"

And of course, "The people answering your question are probably NOT part of the project staff. They're participants just like you. They may be wrong... but they may also be right. Especially if three or four of them are telling you the same thing. And yelling at them about what a lousy thing 'x' is, is pointless. They didn't write it, they can't change it, they're just trying to help you work around it. So be nice. Save your yelling for your email to the project staff. They get paid for it."

I can probably think of a lot more (all it would take is spending a few minutes on the SETI boards...) but I'll leave that up to somebody else. I've got to finish helping a guy on another project that won't hit a key unless I specifically tell him to...

---

Thanks Bill and River~~ and Paul D. and dgnuff and other people I've forgotten, for answering people's questions during this trying time. I guess Distributed Assisstance is a part of what makes DC possible.