Slashdot Mirror
New Seti@Home Client to be Open to Other Projects
An anonymous reader writes "Seti@home is preparing to make a major change to their client and backend. The new system "boinc"
will be a general purpose client and accept work units from other projects (selected by the user).
This will open-up Seti@Home's millions strong user base to academic projects that cannot afford supercomputers. As boinc is an open source framework other distributed projects (think!, folding@home etc) will also be able to use it giving boinc a larger installed base than Seti@Home."
Can I use it to find intelligent life in my apartment complex? The neighbors are keeping me up all night with their parties...
Scientific progress goes "BOINC"?
"They redundantly repeated themselves over and over again incessantly without end ad infinitum" -- ibid.
Will this new client allow for platform specific optimizations? For instance, the RC client took advantage of Altivec which allowed for Macs to absolutely dominate the small computer benchmarks in those ranking whereas they did not perform nearly as well in the SETI rankings. And just so the Wintel weenies don't feel left out and flame me, other platform specific optimizations could also be taken advantage of for Pentium specific calls or even SGI specific calls.
Visit Jonesblog and say hello.
Do you know why? United Devices filed a law suit over some bullshit IP non-sense. BOINC did not have the $ to fight so they had to give in. One of the stipulations was that it could not be used for commercial purposes for the next 18 months. I tried to find the page on BOINC's web site that had the lawsuit info but can't. Rest assured that they will make it OpenSource (OSI approved) as soon as possible.
BOINC is not a Distributed Computing Program. It is a architcture for running DC apps. Good crypto will be used to ensure that a system (server) gets clean data and clients only run apps from that server. You the user will select what DC project you will run on BOINC.
Really all boinc does is help reduce development time for DC projects by establishing a common framework to work within. Someone could run a "Build a better Smallpox program" to build a super Bucket-O-Death (tm) and advertise it a traveling salesman NP hard app to help the girlscouts sell cookies more efficently. There are no safegaurds (AFAIK) on that type of No-NO use.
Is mankind ready for this type of supercomputer (UltraComputer? Hypercomputer?) Seti@Home already blows away all other supercomputers on the planet (I think by at least 1 order of magnitude or so I was told), now with all these different DC projects runnning under the same framework things should get interesting.
Perhaps the IETF will formalise a protocol for DC and take the next step toward a global grid processing system. Think Jabber protocol turned RFC proposal/standards track.