Wales Waves Wikia Search

 
 
By Clint Boulton  |  Posted 2008-01-07
 
 
 
What bigger way to kick off a new year than by announcing that you are creating an open-source, socially driven alternative to search engines from Google, Yahoo and Microsoft?

Apparently, that's what Wikia co-founder Jimmy Wales had it mind with Wikia Search, a new platform the Wikipedia creator launched Jan. 7.

Built from Wikia's July 27 purchase of the Grub distributed search crawler from LookSmart, Wikia Search is a search application like any other, returning links to content and pictures. However, this engine includes social collaboration tools that let users weigh and rank search results.

Moreover, Wikia Search's algorithms are 100 percent open to be studied by the public, providing a transparency beyond Google and others. Wales hopes this differentiation might help the product gain traction among users tired of algorithm changes on the leading search engine.

Wales is realistic; he knows this isn't going to happen overnight. His thinking is five years or more into the future, but he hopes Wikia Search will make a mark sooner. Wales told eWEEK his plans, punctuated by courthouse analogies and visions of caged dining in restaurants, in a phone interview to discuss the product Jan. 3.

How does Wikia Search work?

People can sign up in a very typical social network type of environment. People can enter their interests with various keywords, and you can match people to keywords to find people interested in the same topic as part of the social environment. We have what we're calling the mini-article feature, which allows people to come in, and, in a very wiki-like manner, they can edit the top section of the search results page to be whatever they want. For the algorithmic search results, there will be modern, sexy feedback mechanism where people can rank the relevance of the URLs returned from one to five and then that data can be fed back into the search algorithm as a part of the review process.

A big part of my philosophy that I learned with the failure of Nupedia, which was my first encyclopedia, and then the success of Wikipedia was that I tried to strongly avoid a priori thinking about what kind of tools people need or want, or what they are going to do. It's about short cycles of revising. We have some general ideas about what the rankings are going to look like. We're really in a wait-and-see mode until we've got some people doing good things and bad things we don't want to anticipate.

I use the analogy of the design of restaurant. If you design a restaurant, and you think, I'm going to serve steak, and I'm going to have to have knives, and whenever you give people knives they are going to stab each other, so let's put all of the tables in cages so people can't hurt each other. Clearly, this would make for a bad society. That's not the way we want to live. Unfortunately, that is the way a lot of people design Web sites. They think of all of the horrible things and say, "Well, if you let people rate, what about bots? What about spammers?" I say let's just wait and see. If we see a problem, then we'll think of the least intrusive way of trying to slice that out.

You've publicly stated your dislike for the guarded way Google goes about search. What do you mean and will people really care about how a search is arrived at? It seems more suited for the search engine marketer than the average consumer, who might not care how a search was arrived at.

That's right. There are parts the public will care about and participate in and parts they won't care about. The idea of using open source is a political statement about what we should expect as consumers on the Internet. Part of it is to say this kind of transparency is important. The analogy I would use is that you have a legal right to go down to the courthouse and watch the functioning of the court. Mostly, we don't do that, but it's important to democracy that that is a public process and that we're not having secrets. Even if I don't do that, the fact that I have a right to is a check and a balance.

It's the same kind of thing here. I use the Firefox Web browser, and it's open source. I've never looked at the source code, nor do I care to. I don't know about Web browsers per se and don't really care, but it makes me happy to know that you can go in there and verify what is going on. It's a good thing for me in the same way the courts are public.

Google and other search vendors say they guard their algorithms for competitive reasons, and to keep people from "gaming" the system to improve their search rankings. Aren't you concerned about this?

It is a concern. At the same time, if you talk to people in computer security they'll tell you security through obscurity is a bad idea and that you don't really get strong security out of that. The other thing is, that is a problem you have when you don't have sufficient human input and oversight into a process because then you have to rely too heavily on pure algorithms, which can be tripped up by people gaming the system. If you've got editorial control, that's a different matter and then you've got a different point of weaknesses. What if a community becomes corrupt? Those are interesting social issues, which are much more in my range of understanding.

What sort of privacy policies are you going to have in Wikia Search to keep people secure?

In the social search part of the site, we're trying to make sure every piece of information has a suitable privacy setting that people can adjust, similar to Facebook. People really like it when they can control what kind of information they are sharing or not sharing. We have no plans to move into this idea that we're going to save your entire search history and customize search results for you. I'm not saying that would be something we would never do, but if we ever did it, it would be an opt-in thing. I personally don't think that is particularly useful or is a technological path that is going to work out very well.

In the short run, our policy on the data is that we will run some statistics on each day's access logs so that we know the most popular search terms, and we will be publishing a list of popular search terms so the community can make sure they are well tended to, but then we'll destroy all of the data because we really don't care.

Are you going to monetize the Wikia Search with ads? How would you do that? Text ads, display ads, widget ads, things like that?

Initially, we're just going to have some text ads, but unfortunately I don't have any clever ideas about monetization. The standard sorts of things people do are fine with me.

Rocket Fuel