(A draft of) the next step in search engines is here!

Wolfram|Alpha opened its doors to users this weekend. It is being touted as a step beyond conventional search engines. Its creator is Stephen Wolfram (of Mathematica fame), who calls it a computational knowledge engine. Briefly, this means that this engine can take specially prepared data sets and attempt to help you compute and create knowledge. The idea is that you can ask it questions like you would ask questions of a person, and it would be able to help you understand the answers. See this post from Gina at Lifehacker for a pretty good testing of Wolfram|Alpha, and to get some insight on how to use it.

For those who just want to play with a computational knowledge engine, click here to play with Wolfram|Alpha. Click the 'Examples by topic' link on the right menu to get situated. For those who want to better understand what they're about to play with, read on:

Google is a great way to find information and data on the web, there is no doubt about that. But how good is Google at helping us discover and analyze knowledge? First, some definitions to get things moving:

Information: facts provided or learned about something or someone

Knowledge: information or skills acquired by a person through experience or education; the theoretical or practical understanding of a subject

...so a very rudimentary definition of knowledge is information having been applied or analyzed.

So why are we still just searching on information instead of knowledge? The short answer is that computers can't very easily parse much of the information on the web and make meaning out of it. This is because most of the information is written in a way that is easy for humans to understand.

Could you imagine a world where we slept through the night as our computers chugged hard to create knowledge out of all of our data? This is part of the idea of a semantic web!

As much as Web 2.0 gave rise to a sharper focus on the use of semantic markup on the web, it never quite reached the goal of a fully semantic web. The rise of the semantic web is one piece of Web 3.0.

In my short (several hour) play-test of the engine, I found it to be helpful when I asked questions about the things that the engine was prepared to answer. I found comprehensive answers when using its suggested queries, but I was often left unhappy with the answers it gave to questions that were shot from the hip. For example, the suggested query of 'microsoft vs. apple' yielded considerable results, but the query 'mac vs. pc' yielded no results. This is acceptable to me because I understand that this is an emerging draft, and it can only call upon pre-selected silos of properly scrubbed data.

My thought is that projects like Wolfram|Alpha (and Google Squared, which is dropping later this month) are showcases as to how the semantic web may someday behave. I look forward to the evolution of this kind of software, and to the emergence of a semantic web.