Webware

You can now get a browser search bar for the open-source Wikia Search engine: Wikia Evolution. A search bar is pretty much a requirement to put an engine in front of users' faces, so this is an expected move by the Wikia Search project. But Wikia Evolution moves the concept into new territory.

See previous review: Wikia Search launches the hackable search engine.

Via the Evolution toolbar, Wikia adds a feature to Google and Yahoo searches: It displays rating stars and an "add" button after each result. Clicking on these links adds info into the Wikia Search index, to refine its searches when you use that engine.

Rating Google results for Wikia Search.

It smells. Wikia's Jimmy Wales makes a big deal about how his rating system and open engine are superior to closed systems like Google. Yet this tool asks users to rate Google's own results in order to improve Wikia's. It works, and it's a very clever way to build on the work of others, but it essentially appropriates Google's crawler results for another engine.

I admit this argument would hold more weight if Google and Yahoo allowed users to rate search results. But I believe that eventually they will (beyond current limited experiments), and then we'll have another problem: feedback confusion. How will users determine which service they are rating a result for when they decide to provide feedback?

Mashing sites and content together is the grist for a lot of Web innovation. But taking results from one service and building a direct competitor out of them is not quite kosher, in my book.

Another example of this is Twinkle, which is building its own social network, in part on Twitter; if you to use the Twinkle iPhone app as a Twitter client, you must also get a Twinkle account. And as much as I like it, I feel FriendFeed also skims the cream off the top of the networks that feed it (again, Twitter, but also blogs).

As in evolution, parasites can, over time, strengthen their hosts--or lead to very interesting and robust cooperatives. But they do run the risk of being squashed before they do very much.

(By the way, Wikia Search is really improving. I'm not using the Evolution toolbar, but I have added Wikia Search as an option to my Firefox drop-down list of search boxes. I like it.)

Scour is a search tool that blends results from Google, Yahoo, and MSN together in one stream. You can hot-swap between the three, or break out any single result into another search. If you come across a bad result, or one you think should go above the others, you can also vote it up (or down) or leave a comment--something similar to what Wikia did with its hackable search engine.

On top of the basic layer of search interaction there's also a paying element to the site. Every query you make has a certain point value, with interactions like commenting and voting giving you more points. Once you reach a certain point limit you can then convert points you have into a cash gift card. There's also a referral program where you get a small percentage of the points from the searches your friends do.

I've embedded a demo video below. You can also give it spin here.

This morning, Wikia is rolling out cool features on the controversial Wikia Search engine (previous review).

Starting today, if you do a search on the engine and don't like the results, you'll be able to change them. Your changes will apply not just for yourself, but rather for everybody.

The engine will launch with a smallish subset of machine-indexed pages, about 30 million, which will form the baseline that Wikia Search will let users go to town on. It's "hardly a full crawl" of the Web, admits Wikia (and Wikipedia) co-founder Jimmy Wales, but it's a start.

The editing you can do on Wikia Search is extensive. If you think a result on a search result page is too low or too high in the listings, you can influence its position by rating it. You can delete entries entirely or hand-write new ones. You can also rewrite the text of a search result, including adding code to the result (to insert, perhaps, a site-specific search, like Google's search-within-search).

For people accustomed to the cold (if hidden) logic of purely algorithmic search, these are scary options. It means that your search results are, in part, up to the whims of capricious or crazy humans, or perhaps people trying to game the system to promote some sites while burying competitors.

If you're lucky, though, your search result may be positively influenced by topic experts, your friends, or just other generally well-meaning people. And that's the hope. For the most part, this philosophy works for wikis. Wales obviously thinks it will work for search as well.

I hope it does. I like the idea of an open and transparent search engine. But I'm skeptical, for the sole reason that there's more at stake in search then there is on most wikis. How sites place on search engines has a material impact on how much money they make, so the more successful this engine is, the more people there will be trying to game this system.

To counter this, Wikia Search changes are all done all wiki-style. They're transparent, and they can be reverted by users. Hopefully that will offset the gaming of the system.

One variable that won't influence Wikia Search results: your social network. If your friends rate certain sites higher than the population at large, that fact won't be reflected in the results you get.

Wales said to me, regarding the concept of sorting results individually based on their social network (see Delver), "I'm not convinced that it will be all that useful," but it could be a "piece of data we would use" in the future.

That's probably just as well; the concept of search results directly changeable by users will be weird enough for users to get a handle on.

Wales put a video demo together for Wikia Search. Click to view.

See also: Anoox.

Today, Wikia is planning to launch Wikia Search, a very early-stage version of its open search engine. Wikia (and Wikipedia) co-founder Jimmy Wales believes that it's necessary for the public to take control of search, which he sees as a shared need and thus a shared resource.

The site that's launching today is for users who want to "help us build a search engine," Wales said. So don't expect a Wikia-powered Google killer on Day One. "We want to be sure people aren't expecting a Google-quality experience on Monday."

Here's what Wikia is building: a human-ranked search engine and mini-Wiki, with a social network angle. The first two parts are the most interesting.

Wales says that initially the engine will crawl and index the Web, and give users algorithmically generated results. But users will be able to rank results up or down, which will have a strong influence on further results. This extra, intentional step of rating results will, I am sure, help little-known but high-quality pages rise in the rankings, and encourage link-farmed search spam to sink. Assuming, that is, that users take the time to rate, and that suitable antigaming technologies are put in place.

I tried a preview of the alpha on Saturday and couldn't see how users rate results. Hopefully the rating function will be added before launch.

Google, by the way, also applies human intelligence to its search results; see the Q&A with Google's Peter Norvig in MIT's Technology Review.

Users will be able to contribute mini-articles that live at the top of search pages (much like Ask's search toppers, I imagine). As on any Wiki, people will be able to edit and comment on the articles. This could become the best feature of the Wikia Search experience. These little guides dropped into search results pages by fellow surfers remind me of the hobo code marks: helpful advice left by other travelers who've been before where you are going now.

Wikia Search will have another social angle. Users will be able to find other contributors to work on the search engine with them, behind the scenes from the masses who just want results. It appears that Wikia Search users will also be able to attach their profiles to particular search results, indicating if they are an expert (or, I suppose, have a business interest in) what the search is about.

Wikia Search is all open source. Anyone will be able to download the code for the engine and the crawler. Some people will use this to tweak their Web sites so the Wikia Search engine ranks it more highly. But other people will try to improve the code itself, which could have a very interesting effect on SEO (search engine optimization) industry once and if Wikia Search gets big enough to matter.

Wales maintains that since users will be rating pages, there will be no "search engine optimization" to perform other than creating content and sites people like. Unless I am missing something, even with the option to put human ratings on pages, most pages will remain unrated. The algorithm will still matter. I do have high hopes for Wikia Search, though, since the idea of an open algorithm that is transparent, and transparently updated, strikes me as a lot more fair than the current black-box search engines run by Google, Yahoo, Microsoft, and Ask.

Can Wikia Search succeed? Web search is a difficult and expensive business, but as Wales says, "It monetizes well." Unlike many other crackpot sites we see, with search, if the users come, the money follows. Try out the very early, very rough Wikia Search at alpha.search.wikia.com.

And see also Mahalo, an evolving human-powered Web guide. Currently, Mahalo primarily uses paid staffers to create its topic pages, but a new "Mahalo Follow" feature lets users easily recommend sites to the engine--a more cost-effective way to quickly build a library of human-approved links.

Previous News.com story: Wikia Search to offer first peek next week.