Slashdot Mirror
Citizen Science and Grid Computing
japonicus writes "The Economist has an article summarizing the current state of distributed computing (think SETI@home and its ilk), which suggests that distributed-human projects are going to be the next big thing. (We discussed one such project, the Galaxy Zoo, a few months back.) The distributed-computing platform BOINC is about to expand to human processing. Distributed proofreaders have been a longstanding success (yet inexplicably failed to get even a mention in the article); but there are a lot of other projects waiting in the wings."
I wonder if they could borrow ideas from the wiki community.
That'll have to be first. Not impossible to do, but given the state of the IT infrastructures I've seen unlikely for a while.
Deleted
Hmph..
So basically, -1 troll/offtopic is really slashdots way of saying "I hate that you thought of something before me."
ESP Game:
http://www.espgame.org/
More info:
http://radar.oreilly.com/archives/2006/09/more_on_google_image_labeler.html
Very interesting video:
http://video.google.com/videoplay?docid=-8246463980976635143
"how can they call it a MINE if everything here is THEIRS?!?!" -Straight Jacket
The Economist has an article summarizing the current state of distributed computing (think SETI@home and its ilk), which suggests that distributed-human projects are going to be the next big thing
After all, just look at BotNets. How much more insight do we need than that?
If only Joe Sixpack (who leaves his computer on 24/7 even tho he only uses it about a half hour per day) would understand that every clock cycle is sacred, every clock cycle is great...
If only.
Dedicated Cthulhu Cultist since 4523 BC.
Sure, there are tasks that computers can't do so well at the moment, where giving the work parcels to humans would make the most sense. But can you imagine what micropayments might allow? It would enable a consistent set of trained, motivated workers to be stable over time, and dependable enough to use this kind of network for important activities.
Ultimately, humans get bored and computers don't. But humans can be delayed from boredom quite a bit by financial compensation.
--
Educational microcontroller kits for the digital generation.
The only problem I have with the current way of grid-based computing is that it cost me a decent amount over the year. I have to leave my PC('s) on, which burns up power that could otherwise be saved.
I know several slashdotters leave their computers on 24/7, but I don't. It's akin to leaving a light-bulb on overnight, or leaving the fridge door open. I do have a computer I leave on overnight when it's downloading, but it's a 5headless 00mhz p3 with 256mb ram and it's promptly shut down until I need to download again.
Gone!
SETI was the first to really harness this concept in a useful way, and it was very useful to prove the viability for distributed computing. However, I think the computing power would be better used for medical research, such as Rosetta and World Community Grid. Don't get me wrong, I think the SETI project is valuable, but I think crunching numbers for medical research is more important. I urge all of you crunching numbers for SETI, to also share that crunching power for medical research as well.
The Economist coined that out of their ass? Seriously, the current acception of 'citizen' is a person taken as subject to the laws of a specific government. What does *that* has to do with voluntary distributed computing? Nothing! They just assume voluntary distributed computing = virtuous, virtuous = good citizen, and there bingo citizen becomes synonymous with virtuous. Participation in a common project becomes not a personal contribution, but a contribution from us, *as subjects of a government*.
I'm not nitpicking, this is scary.
\u262D = \u5350
As soon as you see some asshat saying in print or especially on the internet that something is "the next big thing" you can bet your left nut it isn't.
-mcgrew
mcgrew's razor: Never attribute to stupidity that which can be explained by greedy self-interest
You can always pickup a jiggit at http://www.thesheepmarket.com/.
About : http://users.design.ucla.edu/~akoblin/work/thesheepmarket/
Created with : http://www.processing.org/, http://www.mturk.com/mturk/
of the world's biggest Gunrunner
Thanks for your activism.
PatRIOTically yours,
Kilgore
My blog
is listed on my site: http://distributedcomputing.info/ . If you leave your computer on all the time and it isn't doing anything useful when you aren't using it, please look through these projects and pick one or more to contribute to.
Extreme Programming - Redundant Array of Inexpensive Developers
I volunteered for DP for a few months. I got buggy TIFFs that my web browser couldn't deal with, so I sometimes had to work outside the DP proofing environment, which was a pain. (My suggestion that they switch to a more portable format, such as PNG, fell on deaf ears.) And they're still stuck on the idea that plain text is a universal format. That's what made me give up: I was proofing the 1911 Britannica, and realized that a lot of information was getting lost. There was no good way to indicate marginal notes. Both boldface and italic are indicated by all caps. (I REALLY find it hard to enjoy books that are FULL of capitalized words; it DESTROYS a lot of the SUBTLETY. And how do you capitalize "1984"?) And equations were managed with a subset of LaTex which I'm sure I mangled because I didn't have a LaTex interpreter to test it on — in fact, the DP instructions didn't even mention that it was LaTex.
If you want to preserve text for the ages, you have to use some serious markup to indicate things that are part of the content but not part of the linear text. Basically, the solution is to use some form of XML. Yes, I know the arguments: hard to enter, not everybody has an XML browser, etc. There are good solutions that deal with these problems, Just throwing away data in order to keep the document "simple" is not a good solution.
Maybe once these projects pick up speed, people will get together and start working out of a common workplace to increase efficiency !
Hmmm... but then of course they'll need a big building to fit everyone, some form of financing, cubicles....
Hey... wait a minute ! This sounds familiar...
Nope, false alarm. What a new & radical concept ! This could change everything !
A million monkeys and this is the best sig they could come up with...
After all, just look at BotNets. How much more insight do we need than that?
If only Joe Sixpack (who leaves his computer on 24/7 even tho he only uses it about a half hour per day) would understand that every clock cycle is sacred, every clock cycle is great...
If only. That would apply only to sixpackers who don't use MS Windoze which needs to be rebooted daily to fend off BSOD's. I mean really, you're doing some humongus calculation to resolve some combinitorial nightmare of an equation that could revolutionize cancer treatment and just when Joe's PC is about to return the pearls of wisdom that it has generated, your connection blows up because Joe's PC hung because it was Windowized.
I guess this brings up , "What is the MTBF of XP or VISTA"? In a real environment as opposed to a laboratory controlled setting?
This is something that I have had an interest in for the last few years. As such, a large part of my thesis has been developing "CompTorrent". It is a computing platform that has borrowed some ideas from BitTorrent and combined them with distributed computing.
The focus has been on making distributed computing projects as easy to start as a BitTorrent swarm. After spending some quality time with both BOINC and Condor I can assure you that getting a project going from scratch, can be a non-trivial exercise.
Here's a paper if anyone is interested: Enabling grassroots distributed computing with CompTorrent
...a beowulf cluster of these?
Exactly. Amazon's Mechanical Turk system already does this.
http://inchorus.com/ was a startup that tried to do exactly this (think Amazon Mechanical Turk, but distributed). Great fun, lots of projects... no seed funding. :)
http://www.webware.com/8301-1_109-9676000-2.html says what we tried to do.
Don't worry : Joe Sixpack is taking part in distributed computing. Mainly distributed Spamming and distributed DOSing. Thanks to Microsoft's legendary security and modern Zombie worms, all those computer ARE used indeed.
Strom Botnet : brings Grid computing to average Joe's reach (tm).
"Sufficiently advanced satire is indistinguishable from reality." - [Tips: 1DrYakQDKCQ6y52z6QbnkxHXAocMZJE61o ]
Dunno if anyone cares, but I recently saw an ad on some random page that was looking for people to apply for some kind of mass-option trading job (it might not have been options, i'm not good with finance terms). The deal was, you sign up, they give you some software that (presumably) visualizes the fluctuations of various prices of trade-able widgets on "the market" in a chart or graph, give you basic training on when to buy and when to sell, start you off with a bit of dough, then turn you loose and give you a percentage of what you make.
just mentioning it here because someone might find it interesting.
sorry
TANSTAAFL. It's ironic that people running, e.g. ClimatePrediction are simultaneously helping to change the climate. Each PC does not generate much heat, but several million of them certainly do - especially if left on 24 hours a day, 7 days a week as many enthusiasts tend to. See for example this rough analysis: http://hardforum.com/showthread.php?t=1240015
We have to figure both the heat generated and the power consumed (much of which is derived from fossil fuels). Even if you use green electricity, that just means that other people have to use fossil fuels. Good for your conscience, but not for the world as a whole.
On the other hand, as several people have pointed out, the waste heat from PCs does contribute to space heating - thus perhaps reducing the amount of energy spent to keep houses, etc. warm in winter.
I am sure that there are many other solipsists out there.
actually, at least in statistics, is a bit more precise, and you normally disregard data points with more than three deviations of the popualtion mean, as "aberrant" cases. The problem is that normally, random values in a vector do not deviate enough from valid cases to be detectable, so the noise produced by a bot cheating could very well cripple the whole project.
Probably only after a lot of rounds, when tendencies are well known and researched, you could devise more precise tests to check the validity of a given record, but if you know so much of the subject already, your new survey will not be of much help.
This is not a trivial problem in statistics, as normally, statistics are all about looking for interesting deviations from "normal" behaviour. With a simple redundancy test, you take the risk of eliminating the interesting part of the information.
And a corollary of the above, is that as you know more about an object, you can test the validity of observations with more precision, but for this exact same reasson, your results are progresively predictable and less valuable. In other words, the more you know about any problem, the more you determine its outcome, and the less valuable the information produced by new data will be. In the end, everything tends to be normal. or predictable.
entia non sunt multiplicanda praeter necessitatem
I strongly believe the power of volunteer computing project will revolutionize science and technology. That is why I spend all my free time learning and developing a project that attempts to engineer clean energy technology. My work attempts to identify efficient catalysts in hydrogen production using Quantum Monte Carlo and Docking simulations. There is a great deal of developement ahead of me and I work on it with a shoe string budget since it is a hobby. Thanks to the friendly community I am actually making progress, slowly but surely. There are really many more computational problems than projects currently available. Anyone with a little imagination, strong science background and experience developing webservers can start their own! I may think Hydrogen is the best approach to developing clean alternatives. You might think that is a bad idea and have an idea about engineer celluostic ethanol. I would applaud that because in the end of the day Society needs the best solution to Global Climate change. I hope I have inspired people to participate in BOINC. May our community continue growing. Through Meetup.com I started a group that meets in Washington DC to discuss and promote BOINC. Kind Regards, Jack Shultz http://hydrogenathome.org/
http://www.oklo.org/
If you are interested in the search for extrasolar planets, you can download a Java app. It's not distributed computing, though; it's more like a little toolkit where(after setting up the strongest data by interpreting a graph) you tweak some sliders and then tell the computer to integrate and leave it on all night(it's trying to get the best fit to the transit data we have for the star system)
When you get a good result, it's pretty cool. But it takes a lot more effort than the distributed, leave-on-anytime, ones.
I leave my 5 PCs on 24x7. It uses electricity alright, but it cuts down a little on my heating bill, so it mostly evens out. During winter, that is. In the summer this place is so hot... If only there was a way to make them emit cold air instead of hot...