The Daily Cow

Last week, Joe ("Overlord" on the DailyCow.org) provided suggestive evidence that FreeRice may be a scam, but that does not amount to a full-blown proof that it is one. This week (particularly this Friday), through an organized and collective effort, I seek to provide that conclusive proof that we all desire. The question is how, and I've got the answer.

The Premises

The assumption is that a word submission is correct when FreeRice.com says it is--meaning the correct submission is at least acknowledged by FreeRice.com. If FreeRice chooses not to honor that submission, despite acknowledging its validity, it is a lie, a sham; if it does, we should see numbers shoot up significantly (which has not happened yet).

The Theory

Chances are, FreeRice.com does donate a certain amount of money for the WFP to purchase rice with, but likely not in the volume suggested by the number of correct word submissions it receives. There is absolutely no transparency in FreeRice's internal workings or finances, and it has no obligation to let there be any--after all, it isn't even a registered non-profit (indicating that FreeRice.com and its parent website Poverty.com are driven by commercial or business interests); hence, it has the freedom to donate 5,000 dollars of rice when, by its own assertions, it should donate 50,000. Therefore, the WFP recognizing FreeRice.com's donations does not immediately invalidate the argument that FreeRice is a sham.

How We Can Prove It

Working off our premise stated above, I have created a new version that explicitly tests each line of output from FreeRice.com for the keyword that tells us our submission has been acknowledged. It has, by default, 100 "pseudo-threads" (more on that later), to compensate for the extra network time involved. The goal is to get as many people as possible to run this application (downloads and instructions are in that link) in one 24-hour period--I've designated Feb 1 (Friday) as our official day--so we can determine with utter certainty the legitimacy of this commercial exploit.

Minor Technical Details

So you're an EECS major at U.C. Berkeley and you can't believe that I've just bastardized the sacred term threads. Yes, I admit, unfortunately, it doesn't really create that many threads, but it does facilitate parallel-processing much more than does the original version. The single-"thread" version released earlier a week or two ago only sent the requests to the FreeRice.com server and never waited for the output--that would have taken too long since it must take at least 200-500 milliseconds for FreeRice.com to process the data and render an HTML output, and for your system to download and process each page. (For those wondering, that doesn't mean the requests weren't received though; I've confirmed that all HTTP requests are honored even if the "requester" cancels the request immediately after sending it, so the double-check is just to make sure FreeRice's internal timer doesn't affect things)

So, where is the parallelism? It isn't within your computer; it's between the server and you. While the FreeRice.com servers are processing the requests, the application sends another one--and it keeps doing that for 100 times in total. When all 100 requests have been sent, chances are the first request sent will be ready for retrieving and processing. Hence, this construction allows your computer to process more rice as the server works on the earlier requests. Additionally, if you have the bandwidth to spare (the "estimator" is nowhere near accurate, so don't trust it), you can definitely open more than one instance (try not to open more than five, though), and those separate instances will run simultaneously (That configuration is closer to true parallelism on your computer than the implementation I provide). True, the speed at which each instance functions will drop slightly depending on your bandwidth, as expected, but enough to negate the effects of having multiple simultaneous requests.

Get Everyone Involved

Be sure to invite your friends, relatives, classmates, maybe even teachers to participate in this one-day event so we can finally know for sure whether we're being deceived by FreeRice's innocuous-looking message; my theory is we are, and it's easy to prove it. We just need to work together to do this. Don't forget to Digg this campaign.