This is really addressed to the whole BOINC development team, but I suppose to 
David in particular.

I wonder if it would be possible for you to post, hopefully in the 
not-too-distant future, a summary of where we're up to with the whole 
'CreditNew' bundle (which of course includes important functional code like 
per-app quotas and DCF, as well as the credit stuff).

http://boinc.berkeley.edu/trac/wiki/SoftwareDevelopment still has it as a 
"proposal"
We know that a publicly announced Beta test started on the SETI Beta site on 
1st. April 2010. The only reference to the outcome of this test that I've seen, 
apart from my own reports, is changeset 21727.
And it appears, although without announcement, that the whole bundle went live 
on the Main SETI project from Monday 7 June, when it became apparent that the 
new NVidia-supplied CUDA Fermi application needed to be added to the installed 
versions list.

There's some very wise advice on "Recruiting and retaining volunteers" in the 
Wiki, and I'm fond of quoting it:

  "Take an active role in your web site's message boards. Read them frequently, 
and respond quickly to any negative threads that arise. Make a periodic posting 
giving 'insider info' on your project."

In the absence of project staff or developers able and willing to carry out 
that role, and with the SETI forum moderators otherwise engaged, it falls to 
ordinary volunteers to attempt to answer the myriad of questions that arise on 
the message boards - in the first ten days after deployment, I wrote almost 200 
posts, almost all technical in nature, some lengthy, and others requiring 
research and testing on my own computers before I had an answer I felt 
confident in passing on. That's OK: I'm a volunteer, I knew what I was letting 
myself in for, and I'm not complaining. It keeps the brain from atrophying.

But the advice given to other volunteers and the general public on the message 
boards would be more comprehensive, and more accurate, if there could be some 
ongoing background briefing from the centre. Even if those were only terse, 
technical sitreps, volunteers could have a go at expanding them into more 
generally accessible language for the wider readership.

I'm raising this now because the unfinished, provisional BOINC code is 
beginning to spill over and have a detrimental effect on other projects.

Take CPDN, for example. Around 18 November last year, according to Milo, David 
asked for the server code at CPDN to be updated "to enable his script to 
compare the old and new style of credit calculation to run." This turned out to 
be a larger and more difficult job than anticipated, involving much 
customisation, but it's been completed now. Obviously, because the requested 
test included 'CreditNew' issues, Milo didn't take the 'Server Stable' branch, 
but the trunk revision (as would other projects needing support for recent 
developments in GPU and multithreading technology).

According to the latest information in my possession, the Main CPDN project is 
currently at Changeset 21675 - and it shows. Even experienced moderators (and 
CPDN has some of the best in the business, with in-depth project knowledge even 
predating the conversion to BOINC) cannot explain "(reached daily quota of 3 
tasks)" messages, when no tasks have been downloaded for the host, and day, in 
question.

I'd like to think that Milo and other project administrators could find out 
when it'll be safe, and necessary, to update their prohects again to avoid 
glitches like this, without having to rely on seond-hand reports from the 
maelstrom of rumour and speculation that is the SETI message board.
_______________________________________________
boinc_dev mailing list
[email protected]
http://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev
To unsubscribe, visit the above URL and
(near bottom of page) enter your email address.

Reply via email to