Comments on the State of Search/The New Wolfram Alpha Engine

Search technology is of interest to anyone on the net these days, but it has a special kind of interest for those of us concerned with accessing and managing resources for learning. And, for those of us instructional technologists involved in work related to school media centers (or libraries in general), facilitating search is one of the most conspicuous pieces of our day-to-day mission.

I do not claim overwhelming credentials that qualify me to discuss this topic, but over the years I have observed approaches to search, search tools, and alternatives to search emerging and evolving. Today I want to briefly highlight several recent search contenders and then focus on an item launched last week, that is much in the news at the moment, that is intimately related to search but claims to be no search engine at all.

The first four items I want to mention are recent offerings that aim to offer a meaningful alternative to Google. They are:

Cuil ( www.cuil.com )
Viewzi (www.viewzi.com )
Kosmix (www.kosmix.com )
Leapfish (www.leapfish.com )
(And there are probably other worthy items I have failed to include.)

Cuil claims to search three times more web pages than Google (and 10X more than Microsoft). Among other features and claims, all of these package results into user-friendly categories, offer alternative options of how to view results, and generate links to related topics. To one degree or another, these search tools are attempting to generate sort of a "your complete guide to this topic" in a single web page. In addition, Leapfish has added real-time search-as-you-type, or click-free search, which is pretty amazing to see in a web-based search engine.

You can generally find multiple demos and reviews of these tools on YouTube. As an example, here is a video demo of Cuil, which is probably the one that attracts me the most so far:

http://www.youtube.com/watch?v=IMHI3XSg75s


Now enter Wolfram Alpha, the engine that does something other than search
( www.wolframalpha.com ). Wolfram Alpha is a computational engine that specializes not in finding existing web pages but in mining numerical data, especially math and science data, and computing/compiling answers to queries. The "10+ trillion pieces of data, 50,000+ types of algorithms and models, and linguistic capabilities for 1000+ domains" has all been vetted by humans, so the quality of results is expected to be at a reliability level far exceeding what search engines do.

Thus any information that can be expressed quantitatively should be fair game to request.

The problem with Wolfram Alpha, so far, is that people are having very limited success in getting it to do something useful for them. There seem to be two main causes of this problem: one is that users tend to enter queries that are more appropriate for search engines, rather than making the adjustment to the idea of requesting quantifiable information; the other is that Wolfram Alpha is still developing its data resources and there are certainly many gaps that remain in the categories of data that have been indexed.

I think we can expect great things from this new tool, while we can also expect to often see the following result in the meantime:

"Wolfram|Alpha isn't sure what to do with your input."

There is a very helpful CNet video comparing Wolfram Alpha to Google here:

http://www.youtube.com/watch?v=h3emFpcOfek


There is also a two-part introductory video by Stephen Wolfram, who is the force behind the project:

http://www.youtube.com/watch?v=riQ5tpHc_b8


Finally, just my little disclaimer that while information is power, information is not learning. I hope someone finds this discussion of new tools helpful; but there will always be the accompanying question "how do I use this to support meaningful learning?". I'll cop out here for the moment and save that for another day.

Best to everyone!

Greg