The big search engine news story this week was the launch of the new Wolfram Alpha search site, by some hyped to be the new Google. Pandia is not convinced that Wolfram Alpha will make a major impact.

It isn’t. Wolfram Alpha is not an alternative to Google or any of the other major search engines; it is another type of search tool that complements Google. They themselves call it “a computational knowledge engine.”

The name refers to the founder, former particle physicist Steven Wolfram.

In short Wolfram Alpha is a fact finder that is especially fond of statistics. Ask it for the distance to the moon and it will tell you not only the average distance (385000 km), but also the distance right now (363389 km).

If you ask it about how to market your site online, however, it is totally at loss: “Wolfram|Alpha isn’t sure what to do with your input.”

Accessing the hidden web

Wolfram Alpha puts it this way: “Our goal is to build on the achievements of science and other systematizations of knowledge to provide a single source that can be relied on by everyone for definitive answers to factual queries.”

Wolfram Alpha does not spider the web, as other search engines do. Instead itt accesses the hidden web, carefully selected databases that may provide reliable data.

Does it work?

Indeed it does, and Wolfram Research has definitely added a valuable tool our online toolbox.

Still, we doubt that Wolfram Alpha will become very popular among regular searchers, as there are are some problems.

Mostly for the specialists

It is not surprising that it is oriented towards the natural sciences, given that the founder is a physicist. But for those of us who is more oriented towards understanding sociocultural phenomena, this makes Alpha less useful.

Associated Press reporter Brian Bergstein calls Wolfram Alpha a niche Web site for scientists and technologists:”There aren’t many ways everyday Web users would benefit from using it over other resources.”

He points to the fact that it is — for instance — pretty useless for sports, an area where non-scientists look for statistical data:

I kid you not, the query “Super Bowl scores” yielded the response “Wolfram|Alpha does not yet support Romanian.” If you seek a baseball team’s pitching stats, you get a useless chart showing, among other figures, the number of batters a team faced in a season. This computing engine doesn’t compute the earned run average. Obviously, it’s not a golfer.

Finding the sources

The data you get is given independently of any context or interpretation.

You may to a certain extent track down sources (by clicking on the link Source Information), but in the case of the moon question, one of the sources named is “Wolfram Alpha curated data”, which does not help us much in understanding where the numbers come from.

Another link, “Wolfram Mathematica Astronomical Data”, gives us access to a list of illustrations, including graphics of solar system orbit paths, which is useful, but again, it is hard to ascertain from where Wolfram Alpha gets the raw data.

You might say that this does not matter for simple queries, and to a certain extent it does not as long as Wolfram Alpha manages to build a reputation of scholarly reliability, but for some researchers this may be insufficient. But hey, this is a tool for the general public, and not for the scientists, right?

The problem of interpretation

Maybe, but the general public needs help to interpret the numbers.

I have worked with research and innovation policy statistics for many years and know that if people use the numbers without understanding how to interpret them, they may make some serious mistakes.

When I was a kid, and asked too many questions, my parents pointed me to our encyclopedia. The truly believed that whatever the encyclopedia said was the truth and nothing but the truth. It was written by experts, right?

Many years later I have come to the conclusion that there are no such things as facts, not even in the natural sciences. I know too many experts personally to believe otherways.

For most of the data given by encyclopedias there will be a number of competing schools disagreeing on what the metric means. Moreover, it is in the nature of science to change its theories, methods and — even — its basic understanding of what nature is. Which means that you may even question the way the data has been generated.

When you go over to the social sciences and the humanities this becomes even more apparent: The complexity of social and cultural systems is so great that any number you may get out of it is bound to be nothing more than a helpful indicator, not a figure that tells you something essential about the world as it is.

Wolfram Alpha seems to live in a world where it is still possible to tell objective “truths” about the world, which is why it tries to give you the data directly instead of giving you a link to another site, like Google.

Ironically, this is also what stops it from giving you the context. When I do research for a topic I want to have different competing sources. I want to find sites that gives me the story behind the numbers, the background, the theory.

Search engines on a fact finding mission

It should be noted that also the traditional search engines try to give searchers relevant data directly for some questions, exactly the types of questions that is well suited for plain number delivery.

Ask Google to “convert 150 NOK to EUR” and it will put the answer:

150 Norwegian kroner = 16.9194262 Euros

above the first result.

Ask Yahoo! about the population of Norway and it will give you the following shortcut at the top of the search result page:

Population of Norway: 4,644,457 (July 2008 est.) - The World Factbook (with a link to the factbook).

And given all the buzz about Wolfram Alpha, we guess there will be more of the sorts.

In other words: Wolfram Alpha’s competitive advantage may be short lived, unless — of course — it manages to sell its data to one of the big search engines. That may be one very efficient way of monetizing the Wolfram databases and technology.

How to ask questions

Interestingly enough, Wolfram Alpha does not understand my query “R&D investments in the USA”, probably because the answer belongs to the social sciences and not to the “hard” natural sciences. There are databases it could have made use of to answer that question, though.

Or maybe I am not asking it in the right way. Wolfram Alpha has difficulties in understanding more natural like language, and I afraid most users will not have the patience to learn Alpha’s own way of “asking”.

Herbert Roitblat, CEO of Truevert, puts it this way over at Alt Search Engines:

At this point, Alpha’s natural language understanding is also quite limited. For example, if you type in “difference in calories between a 100g of beef and 100 g of turkey ” you get nothing, though it does “know” about the caloric content of these two foods. Typing “calories 100 g beef - calories 100 g turkey,” also seems to confuse it. On the other hand, entering “calories McDonald’s big mac - calories Burger King Whopper” returns the “right” answer, -270.

To sum up:

Wolfram Alpha will be a useful tool if you are looking for numbers in particular disciplines or areas, especially within hard science and technology.

It does not help you gain an understanding of what the data means, though, and in many ways the regular search engines may help you find the same data in a sufficiently helpful way.

We will come back with an article on how to use Wolfram Alpha for particular types of queries.

See also:
Wolfram|Alpha - computational is the key word! Phil Bradley
Five Things Wolfram Alpha Does Better (And Vastly Different) Than Google Mashable
Wolfram Alpha Isn’t Google, so Stop Comparing Them Fast Company