Last week after much pre-release hype and fanfare, Wolfram Alpha was finally released. What it can do, it does very well, but as I suspected, there are limitations because of its database of information. For a fun test, I entered the following search (I actually saw it another article):

How many roads must a man walk down before you call him a man?

The Wolfram Alpha database programmers must be Dylan fans because it gives the result:

The answer is blowing in the wind (according to Bob Dylan).

But if you ask:

How many ways are there to lose your lover?

Wolfram Alpha doesn't know because apparently the programmers aren't Paul Simon fans, who once wrote "there must be 50 ways to lose your lover."

This is admittedly a frivolous example, but it does prove that the answers you get depend on what's in the database.

Frankly, I found it hard to get any answers out this search engine at all, no matter what I asked. Maybe I lack the analytical mind that the underlying Mathematica engine requires. Sure, you can get data like the weather on the day you were born, or how many days old you are, but you have to know what to ask and how to ask it. What's more, if you ask something the engine doesn't know--something I had an amazing knack for doing--you come up empty.

For a certain narrow set of users--researchers, librarians and students--Wolfram Alpha will provide a solid tool, but it's really only one tool in an arsenal of search tools, and not the last best search engine you will ever use. Google's not going anywhere.