[an error occurred while processing this directive]

learning.now: at the crossroads of Internet culture & education with host Andy Carvin

[an error occurred while processing this directive]
March052008

Human-Powered Search: Just What the Teacher Ordered?

Educators have lamented the quality of search results since the invention of the first search engine. All too often the results pages are littered with links that are useless in the classroom - or worse. But what if those search results pages were generated by people - educators and students in particular - rather than by complex algorithms?

Using search engines has become such a mundane part of our lives, we rarely give them much thought. It’s as consistent as gravity in that way. You just enter some word or phrase, and the search engine will crank out some kind of result. They may not be the results you’re looking for, but you’ll still get something. (Here’s a fun game - try to come up with words or phrases that generate exactly two results. Makes for a great ice-breaker on blind dates and the like.)

We use these tools every day, but most of us don’t consider what’s going on under the hood. Typically, search engines are powered by complex mathematical algorithms that take all sorts of factors under consideration - relevance of keywords and metatags, popularity of a site, the number and intensity of other sites linking to it, etc. Different sites may use different algorithms, but they’re still algorithms. Powerful servers do the crunching as soon as we hit the submit button, expecting instantaneous results.

But now there’s another type of search: one that’s human powered. It’s just what it sounds like: actual human beings spending their time determining what are the best search results for different words. So when you search for a particular word or a phrase on a human-powered search engine, the results you see were determined by some person or persons who concluded that these are the results you were looking for. It’s not exactly an instantly scalable approach to search, but proponents argue that adding the human element increases the relevance of the results while also cutting the number of spam links.

One of the best known examples of human-powered search is Mahalo. Launched last spring by serial entrepreneur Jason Calacanis, Mahalo invites users to suggest links for specific search results, then has a team of editorial staff vet them to make sure they’re neither spam nor off topic. While Mahalo has started to generate interesting search results for topics ranging from technology to pop culture, it hasn’t progressed far when it comes to education topics. Searching for the word education only leads to four results, while “education sites” generates just a single hit. And don’t even try searching for Seymour Papert, as it suggests links to Philip Seymour Hoffman and Jane Seymour instead.

Meanwhile, there’s Wikia Search, which launched in January. I mentioned Wikia in my last blog post; it’s the commercial wiki service launched by Wikipedia founder Jimmy Wales. The site is now experimenting with an open source version of human-powered search. Rather than employing paid staff to edit search results, as Mahalo does, Wikia search results are created by anyone who wishes to edit them, just like Wikipedia. The tool, however, hasn’t exactly gotten stellar reviews. A review published by Techcrunch declared “Wikia Search is a Complete Letdown, while searchengineland quipped, “it’s really just yet another crappy search service that may, potentially, if all goes well, eventually turn into something useful.”

Now some of you may be wondering, why is Andy telling me about sites that don’t seem useful? I’m so glad you asked. Frankly, I don’t know if they will be useful for you or not. So far, they haven’t exactly become destination sites for me personally. And I’m not completely sold on the idea that human-powered search will eventually be better at parsing my search goals than algorithm-powered search engines.

Having said that, there’s something about the idea that fascinates me when examined in an educational context. Over the years there have been many attempts to create directories of useful educational content for students and teachers. It’s been over a decade since the MCI Foundation’s Marco Polo project, for example, created a coalition of organization and experts to vet sites for inclusion in their various directories. Many other projects have come and gone since then as well. But Mahalo and Wikia search offer two variations of a very different approach to previous efforts. Why employ experts for when the general public has expertise as well?

In the case of Mahalo, I could easily see them or another company reaching out to the education community to get involved in submitting links for search engine results based on the collective Internet knowledge of teachers and students. With that model in mind, there would still be an editorial layer - editors with an educational background, hopefully - that would review the suggestions and organize them in a coherent fashion. But an educational version of Wikia search would eliminate that layer as well, letting the users determine the fate of each search result page. A crowdsourced, open source educational search engine, if you will. (Rolls off the tongue, doesn’t it?) As is the case with Wikipedia, this method would allow for achieving much greater scale more rapidly than professional editors might, but with much greater chaos as well - which could be glorious, maddening or both, depending on your perspective.

For now, I’ll continue to remain somewhat skeptical. But that doesn’t mean I wouldn’t be happy to see an intrepid group of students and teachers try to beef up educational search results on sites like Mahalo and Wikia. Boon or bust, it certainly would be a learning experience. -andy

Filed under : Cool Tools, Websites, Wikis

Responses

Good post Andy. I do have a few thoughts on this issue.

If someone did build an education version of Mahalo or Wikia, would teachers and students be able to use it in the classroom? Or would the admin or tech folks at a school block it like they do so many other useful web 2.0 tools?

For example, both Google and Yahoo! are blocked at my school. So are Flickr, VoiceThread, Yackpack and PBWiki.

How would such a venture comply with CIPA and COPPA requirements? Moreover, if such a thing did exist, would teachers really use it? So many of my peers think Wikipedia is “evil” because anyone can contribute.

While there are lots of great web 2.0 tools out there, very few of them actually get used in the classroom. Most of my colleagues don’t even use email, let alone any fancy web 2.0 tools.

The US school system has a long way to go….

I think Curriki is trying to do this but within the context of curriculum.

Have people tried using the portaportal site? I like this because once you’ve taken the time to set it up and put in your links you can have students access only the sites you allow.It would be great if educators could link to other educator’s portaportal sites. The one draw back is that it is time consuming to set up.

I thought to myself, “Why not check out Mahalo?” I typed in “Women in History + Science” in the search box and rec’d some hits. Out of the 6 I tried, 5 of them were not in use.

Guess I won’t be using this search engine…

I thought to myself, “Why not check out Mahalo?” I typed in “Women in History + Science” in the search box and rec’d some hits. Out of the 6 I tried, 5 of them were not in use.

Guess I won’t be using this search engine…

Interesting post! I agree that human-powered search engines have a lot of growing up to do, but I do think they have their uses right here and now.

Say for instance I want to assign a research project in my Social Studies class in which each group presents on a different country. While a search for “France” on google results in mostly commercial travel agency sites, a search on Mahalo gives me seven educationally-relevent sites, including the CIA World Factbook and results for “France” in BBC News.

Let’s not dismiss Mahalo quite yet!

Andy,

Thanks for taking a look at Mahalo. We’d be glad to have teachers and students submitting links and helping to build out some of these topics.

If you suggest a link to a page we don’t have it will automatically create the page (these are called use generated pages).

In the meantime we are going to take care a few of the terms you mention here.

Sean Percival
(Mahalo guide)

Andy,
Very interesting post, but as Tina stated schools have most of the websites blocked making it very difficult to use. However, I know recently my school library has allowed for students to access specific websites that will provide students with more specific links to the information that they are looking for, similar to Mahalo. I think that human-powered search engines should be available for people to choose to use them, but I think that we should depend solely on them. Like everything else they have their advantages and disadvantages.

Another great software is Jumper 2.0.

It is open source software available on Sourceforge (project jumper) that allows users to apply knowledge tags to information resources on either side of the firewall. Jumper allows you quickly find and share the best possible information and resources in the deep web.

[an error occurred while processing this directive]