February 6, 2002

12:50 PM

With a grand prize of $10,000, the First Annual Google Programming Contest is sure to catch the eye and keen mind of many a programmer who has thought "Oh, what interesting things I could do with 900,000 web pages in a pre-parsed and raw format!" The Google phenomenon has already inspired add-on search tools and a game or two. What new tricks will add value to this handy web repository?
posted by girlhacker (25 comments total)

Neat contest! I'm hoping for someone* to figure out a way to make "find similar pages" get you closer to your goal. Intuitively, it seems to me that finding similar pages show be a way to narrow my search slowly.

Sort of the way that you can train websites to suggest books for you.

* I've got no C++ skills, so I'm releasing this vague idea in the hopes that it finds a good home.
posted by iceberg273 at 1:00 PM on February 6, 2002

googlepersonals! enter the search terms of your choice and get matched to the personal web site owner best suited to you! of course, the whole "feeling lucky" thing might be a bit much...
posted by judith at 1:38 PM on February 6, 2002

I would like something like that, except with online shopping, I would like to be able to go to shoes.google.com, pick my color, style, brand, size, and then get a listing of product pages.
posted by corpse at 1:43 PM on February 6, 2002

Neat contest! I'm hoping for someone* to figure out a way to make "find similar pages" get you closer to your goal. Intuitively, it seems to me that finding similar pages show be a way to narrow my search slowly.

Sort of the way that you can train websites to suggest books for you.

Google already pretty much uses the same algorithm (with modifications, of course) that sites like amazon.com use to suggest books to you. They basicly select users from a database that have looked at the same page you have. They then iterate though this user list, and the select other pages that these users have looked at, and then suggest these pages to you.

The more different pages you look at, it will iterate though the user list looking for people who've looked at the same pages your have. It just continues to narrow it down until it gets a few results.

Of course this is a gross simplification of the process. The fuzzy logic involved would make my head explode. But google works in almost the same way. It lists pages in order of how many pages link to it, and it sorts those pages by how many pages link to them, and so and and so on, recursively.

So, pretty much, google already does what you want it to do :-)

But this is a great contest. If I had some free time, and about 10 more years of algorithim design knowledge and traning, I'd enter.
posted by SweetJesus at 2:04 PM on February 6, 2002

I can't help feeling this is uncomfortably similar to the Limp Bizkit thread. $10,000 seems an awfully cheap way for Google to pick up a lot of new ideas about what to do next with their search engine.

Personally, I'd like the ability for Google to be able to recognise when a search team refers to more than one target. For instance, a search on "Genesis" would clearly seperate between bible searches and music searches.
posted by salmacis at 2:14 PM on February 6, 2002

$10,000 is a bargain for what is google getting out of this deal. i'm a consultant and my company pimps me out at $155/hour. so they're basically paying for 65 hours worth of work at our consulting rates.

and the people that are going to participate are only going to be the grad students in data mining and what not, who are much more valuable than their typical programmer. combined with the fact that all code produced belongs to them, they are paying a pittance for the next big thing.

i'm kinda disappointed with google for this one. i've always respected their business practices. at least offer a job to the winner rather than $10k, and let everyone else keep their ideas. that seems like it would be the more "google" thing to do.
posted by AaRdVarK at 2:30 PM on February 6, 2002

Here's another add-on search tool for Google.
posted by bravada at 2:30 PM on February 6, 2002

Google is not doing collaborative filtering, like Amazon is doing for their customers. They are really doing similarity filtering based on links pointing to a web page. This does *not* take into account the reader's tastes. I think that's the main problem with using Google for fresh content (i.e. news links).

Pardon the self-link, but that's exactly what I am trying to do. Tech news only for now, and a simple filter because there are not enough users, but stay tuned...

In effect, I am trying to build an automated weblog where the content adapts to the user and stays fresher and more relevant than a human-edited weblog. Not that my site is totally automated, but it's getting there.
posted by costas at 2:40 PM on February 6, 2002

AaRDVark: I guess some of the people whose code impress Google, will get job offers from Google. The guy who actually wins the contest will also get a tremendous amount of brand equity in the development community. That is more valuable that actual $ in terms of future employment/VC funding/resume' perspective. If you read the contest rules, it also says that the winner would would give "Google a worldwide, perpetual, fully paid-up, non-exclusive license to make, sell, or use the technology related thereto....". I believe that means that the winner would be free to sell/use the software elsewhere. They are also talking about using GPL products for all development. Doesnt strike me as a very dark plot.

But I guess there would also be many who would write good code and throw up good ideas that Google may possibly mine. Which is probably not very fair. Ah well...we dont live in an ideal world .....
posted by justlooking at 2:51 PM on February 6, 2002

. With regard to an entry you submit as part of the Contest, you grant Google a worldwide, perpetual, fully paid-up, non-exclusive license to make, sell, or use the technology related thereto, including but not limited to the software, algorithms, techniques, concepts, etc., associated with the entry.


Wow, what a scam.
posted by delmoi at 3:13 PM on February 6, 2002

Reminds me of the Limp Bizkit guitarist auditions.
posted by mecran01 at 3:25 PM on February 6, 2002

It is already the case that Google has significant intellectual property rights in any technology which is designed to tap into Google's database, as any such product is a derivative work of Google's pre-existing intellectual property (i.e., its data architecture, user interface, etc.) It would certainly be very difficult for anyone to stop Google from using, without payment of royalties, any meta-Google technologies developed by third parties, whether or not for a contest.

The only real question of Google business ethics that remains is that in encouraging the development of intellectual property to which Google can glom on, by means of a contest with a $10K top prize, is Google somehow not playing fair. I don't think so, and I am happy to let people decide for themselves what's fair or not, in any event.
posted by MattD at 3:37 PM on February 6, 2002

As I pointed out in my previous comment, if you win it, you grant Google Non-exclusive rights to use your code. That means you can give away the same code free under GPL, sell it to every other search engine company, print the code on your t-shirt, doesnt matter.

It is also going to cost Google a significant amount of money to wade thru thousands of lines of lousy code that they are also sure to get and that would probably do something they already know how to do. If you think harvesting IP from all that junk is going to be easy,think again.

I am sure that they would also run into some pretty innovative new ideas. More importantly, they would discover some brilliant kids out there who code for the challange of it and I think that is the more important discovery that they would make.
posted by justlooking at 3:47 PM on February 6, 2002

Uhuh and people complain about unemployment :) Well if Google really wants big improvements why don't they hire a bunch of good programmer after a selection ?

"Nay selection is too expensive" I hear some manager in Google screaming that right now.

Ok then do the contest, but give the 10 best programmers (google be the judge of programming quality) a position in google, not a miserable 10k for a work worth hundred of thousands of dollars if not million of dollars . An illegal immigrant working at the corners of the streets earns more !
posted by elpapacito at 5:05 PM on February 6, 2002

for a work worth hundred of thousands of dollars if not million of dollars

...well, if you're convinced that your idea is worth that much to you then maybe you won't be submitting it for potential $10,000 prizes, eh?

rather than think of this as google's evil plan to screw brilliant programmers out of their hard-earned billions, maybe you could think of it as a way for GPL'd software projects to get a $10,000 grant.
posted by sad_otter at 5:22 PM on February 6, 2002

There are good things about this, but the examples they give are what put me on the side of it being kind of a scam. I mean, "designing and implementing an efficient index structure"? That's the core technology that makes or breaks the entire company. If you can figure out how to do that better than the combined talent employed by Google, you have an idea that should be worth a lot more than $10k.

In fact, if you're a smart kid and you figure out an idea like that, I'd highly suggest forming your own company that demonstrates that the technology works, and then trying to get Google to buy it, instead of giving the idea away to them (as even the $10k isn't guaranteed).

If the desired applications were a little more like minor add-on features than core technologies, I'd feel a little better about it.
posted by mcguirk at 5:47 PM on February 6, 2002

sad_otter: i'd rather first copyright my idea (if i think it's worth millions i'll protect it, but I hear it's expensive) then
send it to google for them to review. If they like, they can buy , if not they have spent 0. If they try to cheat, I'll bring em in court (ooh strange it's bloody expensive as well but I have a good chance to win).

Then after google has bought it, I'll fund any Open
Source stuff I like with a condition: it must be not used for commercial purposes.

If you follow path one, the easy one, google gets a sizable share of money with 0 investment, you get a dime, gpl gets a dime.

If you follow my path, even if it's harder, and you win, google will have an EXCLUSIVE right to use the idea, you'll have much more then a dime, Open Source or GPL will have more money then they could shake a stick at.
posted by elpapacito at 5:47 PM on February 6, 2002

elpapacito: The problem is an algorithm can't be copyrighted. It can be patented, but that's usually pretty complicated and expensive, and you _must_ pay to defend the patent against any and all infringers, otherwise it becomes invalid. Also, outside the scope of this contest, Google is not going to be willing to review every random idea some nut off the street sends in. (In fact, it's dangerous to even look at stuff like that due to people later claiming stolen ideas.)

This contest is kind of like GM coming out and saying, "Design a new fuel injection system for us! If we like it, we'll use it in all our cars, make millions of dollars in profit, and give you $10k!"
posted by mcguirk at 6:13 PM on February 6, 2002

mcguirk: very interesting, thanks for pointing out
posted by elpapacito at 6:48 PM on February 6, 2002

The same concerns were probably raised about AOL Greenhouse when it started up. The project encouraged nobodies with good ideas for online content to pitch their idea for an online content area. AOL funded the best ones, resulting in several successful ventures, including The Motley Fool.
posted by rcade at 8:07 PM on February 6, 2002

you _must_ pay to defend the patent against any and all infringers, otherwise it becomes invalid.

This is incorrect. It's true of trademarks, not of patents. Patents, you can just leave lying there for years if you want, then start enforcing them. That's exactly what Unisys did with their LZW patent, in fact.
posted by kindall at 10:54 PM on February 6, 2002

Ok then do the contest, but give the 10 best programmers (google be the judge of programming quality) a position in google, not a miserable 10k for a work worth hundred of thousands of dollars if not million of dollars .

They may be thinking about doing something like that, but it would be incredibly stupid to publicly promise to hire people with the top 10 entries when they don't know anything about who will respond and how those people's talents might fit with their current needs for programmers.

More likely, they're hoping to find a pool of good programers, of unknown size (5? 10? 50? who knows?) from which they might find a few good prospective employees.
posted by straight at 6:36 AM on February 7, 2002

The slashdot thread offered up some clever ideas on this subject, but the best I saw was an "annoyance" rating.

A script would parse each webpage looking for needless javascript, pop-ups/unders, midi, etc. and rate each page's annoyance factor in your search results. Ready, set, code!
posted by johnnyace at 8:04 AM on February 7, 2002

It would certainly be very difficult for anyone to stop Google from using, without payment of royalties, any meta-Google technologies developed by third parties, whether or not for a contest.

I would be interested in hearing a more detailed fleshing out of this idea.

For instance, one could argue that every application written to run on a Windows operating system is a "meta-Windows" technology, yet Microsoft doesn't charge all software vendors royalties.

Or, to step out of the technology scene altogether, let's say I create and sell automobile cupholders. If I design and sell a cupholder that is created exclusively for a certain make of car (because normal cupholders won't work in that car due to a specific design feature) that I'll call the foomobile, have I not created a meta-foomobile cupholder? Does Foomobile, Inc have the right to charge my company royalties for selling a product based on their original design?

There must be countless real life examples of this scenario, both in and out of the tech arena (I've thought of many while penning this post). As far as I am aware, very few of these scenarios involve the derivate product maker paying royalties to the creator of the original product...
posted by syzygy at 4:15 AM on February 8, 2002

After rereading MattD's post, I'll retract mine. "Anyone form using" is the clincher.

Entschuldigung...
posted by syzygy at 4:29 AM on February 8, 2002