Stratcat wrote:

"Klaatu" <mutster8_nospam_@netscape.net> wrote in message
news:R97Qc.122$uW4.23@newssvr24.news.prodigy.com...

In case you haven't seen.  Dave Anderson's BOINC climateprediction

lecture.

A 31MB davea.wmv file.  Sort of interesting, one comment was that
oveclockers sometimes return bad data.
http://www.climateprediction.net/video/openday/davea.wmv

I caught this yesterday, and dropped in the CPDN forums to discuss
posting a link, but it looks like you beat me to it!

I highly recommend checking it out. A serious BOINCer or BOINC
beta tester will probably know the majority of facts presented, but it's
pretty interesting reading the nuances and what's said 'between the
lines'.

Some things I picked up on:

- Yeah, the OC thing. But IMO, he laid possible discrepencies on
the OC crowd, and completely side-stepped the cheating issues which
were also a driving component for the current 'doublecheck' and 'WU
tied to a machine' design (Understandable, given the audience).

- Noted that team formation led to team competition, which turned into
an effective mechanism to enchance user recruitment.

- Stating that Arecibo, at times, could not produce enough work, and that
WU's were sent out dozens of times, due to the users getting upset when
they have no work (IOW, churning SAH1 WU's beyond 'reasonable'
scientific needs, in order to keep the users content).

- Envisioned BOINC growth to a point where "anyone, w/o any central
authority" could easily and inexpensively bring a project public,
via BOINC.

- Stated BOINC could be used for "video rendering".  <Yikes!>

- Stated DC has <1% consumer PC market share, but (he?) would
like to see this improve to 10 - 20, or even 50%.

- Stated he envisioned a 'BOINC like' system (grid computing?) before
SAH, and saw SAH as a vehicle (opportunity) to impliment his idea(s).

- Briefly touched on his association w/United Devices, though not by
name.

- He got a pretty hard chuckle from the audience when he stated BOINC
was 'still developing'(par) during the last month since it went public
(SAH2
BOINC's  start-up issues have been discussed in the CPDN forums).

+++

Lotsa' insight you just can't get from a dry 'paper' that comes thru in
verbal delivery and body language.

BTW - Lasted ~30 mins, and worked fine using WMV, w/o a d/l, as
a direct connection using DSL. The CPDN server b/w was excellent.

Good summary, the BOINC and related history was interesting also.