In Search of the Next Search Engine

In my unceasing efforts to stay current I inevitably tried out the search engine du jour, even though I have very little criticism of Google compared to past search engines. This month's much-hyped entrant into the race was Wolfram Alpha, from Stephen Wolfram at Wolfram Research. The concept, as Stephen explains it on his blog, is that, finally, we will be able to search not just for the whereabouts of a piece of knowledge, such as a list of finalists on American Idol, but the ability to search for an answer. I would question his statement that we have never been able to do this; in fact that is exactly what the first computers did, allow people to enter an algorithm by way of a program (and yes I did used to mess with a thousand punch cards, praying that the operator would not drop them!). As Alan Turing, the father of computer science famously said "A computer would deserve to be called intelligent if it could deceive a human into believing that it was human." I believe that Mr. Wolfram is referring to the average user who does not wish to learn a computer language or boolean mathematics in order to ask a question of their PC.

There are two major flaws to Wolfram Alpha. First, the database that it draws on is very limited. This is not a search engine of the internet, rather it relies on data from a dozen sources. That's it. In fact, when you look up the source, it is usually listed as Wolfram Alpha curated data. The US Census data, the CIA World Factbook, the Times Gazetteer, all these are accessible by the user and give far more in-depth and varied responses than those of Wolfram. I spent some time trying out variations of the suggested examples; names, dates and cities. David Green returned the information that David is the 818^th most popular surname in America (all name information is derived from the U.S. Census). Apparently you cannot find out about a combination name, i.e. you can compare Smith to Green or David to William but I had to look elsewhere for how many people in the world have the same name as me. Incidentally, when I tried to be a little more definitive, asking for David (given name) + Green (surname) the search engine was flummoxed. Should you wish to have a ranking of popular names then you should go elsewhere, Wolfram doesn't do rankings; however, a 10 second search via Google will give you rankings for the past 150 years.

A search for London and New York yielded the current population of both, the current time and the distance apart with approximate flight time. It did suggest I look up London and New York on Wikipedia; now that was a good idea, why didn't I think of that? I was born in London so I decided to look up historic populations such as 1900. No luck there either; the sources, once again are the U.S. Census or the Columbia Gazetteer and they only have the current data. Moving on to socio-economic data we find much the same story. The GDP of Spain is $1.47 trillion, how this compares to other countries we do not know and cannot find out short of laboriously entering country after country and jotting down the answers so that we can do our own ranking. OR, we could do a Yahoo search for the CIA world Fact Book and look up the relevant table, but then that is so retro, I mean that is the way we used to do it.

Which brings me to the second flaw. Anybody searching for arcane knowledge, writing a thesis or just wondering about different molecular structures would not want this superficial, dull and limited response to his query. Regular users of the internet generally know where to go for specific questions, Wikipedia, the relevant Census Bureau, university websites or, as in the case of another of their suggestions, IMDB if you wish to know anything about movies. Granted the site is improving, ironing out the kinks and broadening it's scope ( a week ago when you looked up any date in history it told you what day of the week it was, as well as the sunrise and sunset on that date for your IP loacation, now it at least lists an event that occurred on that day). Yes, it's a fun toy to play with for a while and I shall make a note to look at it the future, maybe in a month or two, in the mean time I am sure that the guys at Google, Yahoo and Microsoft are not losing any sleep over it, let's see how Bing turns out.

Sources: http://blog.wolfram.com/2009/03/05/wolframalpha-is-coming/
http://www.wolframalpha.com/
http://google.com