Tech Musings

Microsoft has just now released their “new” search engine, Bing. For one, they got the name right this time. Did you “bing” that? You get the picture.

In recent history, highly hyped search engines, aka “Google Killers”, have bitten the internet dust in a sad way. Over the last couple of years or so, Ask, Cuil, Searchme, Cosmix, and now Wolfram Alpha had raised the expectations to great height with marketing and media campaigns (remember “Algorithm”, the marketing campaign of Ask?). When you raise the level of hype the way it has been done in the past, the chances are that people will give you at least one look. You have to look really really good when they do, for they may never look back at you. You never get a second chance to make first impression, as some philosopher once said.

Well, back to Bing. I chose a few queries to test-drive this spanking new package over the good ole Microsoft Live search. So, what do we see?

Query: CMU

This was something on my mind, since my son asked about it last night. Of course, CMU refers to Carnegie Mellon University. And what do we see on Bing?

And on Google?

And on Yahoo?

Well, Google and Yahoo are bang on target. Bing? Putting Central Michigan University above Carnegie-Mellon in search result for CMU? You must be joking!

Query: Tata Nano

Bing gave me some stale news, Wikipedia link, and link to Tata Motors. Google? the same, minus the news bit. Yahoo? took me to tatanano.com - now that is the winner!

Query: who won britain’s got talent 2009?

Bing and Google does better than Yahoo on this, essentially based on number of wiki or answer links. The only one where the name of the winner (Diversity) showed up in the results was Google, so that would be the winner.

More queries

I did some more queries:

“green sand beach big island hawaii”,

“pandu college”,

“best 1080p tv under $1200″,

“php $_server”,

“SFO to BLR”,

“height of Kilimanjaro”,

“dust mite allergy symptoms”,

“Where is Timbuktu”,

and, finally, “Britney Spears” (the ever popular query).

In none of these queries, any search engine consistently outshone the other. They more or less found the same old results, had the same old presentation, and very often needed me to sort through the 10 blue links to find the answers.

We are not yet there with search.

No, not with Bing. Not yet.

Wolfram Alpha is the new golden boy in the block, the new “Google Killer”. There has been a procession of sorts of Google Killers in the last couple of years, from Ask to Cosmix, to Searchme to Quil. So, with some tired anticipation, I took Wolfram Alpha, the Wolrd’s Computational Engine, for a test drive.

First the news: It is not a search engine.

Second, the bad news: It is not a “Google Killer”.

Third, the good news: It is good at what it does.

So, lets get through the test-drive of Wolfram Alpha.

Query: Paul McCartney

Nice: Full name, Date of birth, Place of birth etc. Useful information.
Not so nice: the timeline graph is a dummy.

Query: Betales

Not so nice: The topic is under investigation! Beatles?

Query: Zaheer Khan

Not so nice: No result.

This, as I would later find, is the essence of Wolfram Alpha. It knows about a very few things, and it knows them well. For the very vast amount of the rest that it knows nothing about? well, there is Google.

And you call that a “Google Killer”?

Query: Mono Lake

Nice: 69.5 square miles area, 141 feet deep, elevation 6381 feet, Satellite image (takes you Google maps!).
Not so nice: Map does not work.

Query: Ulsoor Lake

No result. Gives me Lake Ulster, NY, instead. Hmmmm. I wanted the burger, not the frys! When I look at Google, there are 34000 results, including one from Wikipedia.

Query: Bangalore to Miami

Nice: Distance, flight time, populations, elevations
Not so nice: No weather information

Query: 45 mph

Nice: conversions to other units.

Query: C Eb G C

Nice: Finds the music scales.

Query: Height of Mt Shasta

Nice: finds the height.

Query: Height of Mt Tamalpais

Not so nice: Mt Tamalpais? Hmmm. Whats that? We will need to ask Mr Wolfram.

Queries: Crow, Bangalore, Yahoo, Mission College, Thar Desert, Nike, Citric Acid

Nice: Works well, with a lot of category-specific details.

Queries: Length of Honda Accord, Pandu College, Latitude of Thar Desert, Sepoy Mutiny, Nike Air Jordan, Bokeh, Nikon D80

Not so nice: Knows nothing about these.

Well, I guess I am done with my test drive. Wolfram Alpha is very very far from a search engine. It knows about a few things, and it knows them well. It does some computations (presumably using Mathematica) to show some interesting results about the few things it knows about.

I would not write off Wolfram Alpha. It has complemetary technolgy which is likely to be useful to “real” search engines, such as Google, Yahoo and Microsoft.

Wolfram Alpha is, therefore, a good add-on.

And an aquisition target.

Soon, I guess.

I sort of grew up with SGI (Silicon Graphics) computers during my university years and beyond. There was a time, for a span of some 12 years in my academic life, when I had a SGI machine either on my desk, or somewhere close by. Most of my research was done on SGI machines, and data visualized on its fabulous graphics system.

So, today was a sad, sad day for me. With a lot of regret today I read the obituary of SGI. An once mighty Silicon Valley company, with peak sales around 4 billion US dollars, now got sold for a paltry $25 million to Rackable Systems. This is the end of “beautiful computing” as I knew it.

For Silicon Graphics machines were as much work of art as they helped produce. For, do you know, there was a time when the Hollywood dream machine rode fanciful on the able shoulders of SGI. But that was then. Cheap and plentiful hardware and “utility computing” has taken its toll. The sheer power of combined masses of the Intel and AMD, NVidia and ATI, has steamrolled past the fragile beauty of SGI.

I am sure the management of SGI is to be blamed aptly. But I will leave it to others to do that job.

I will only bemoan the beautiful computing machines that have once graced my desktop, and computer rooms, and provided number crunching and graphics horsepower. But, most of all I remember being ecstatic the day I got my brand new SGI Indigo. No, I haven’t felt quite that way with my MacBook Pro, though it comes close at times.

And, no, please don’t even talk to me about the “functionally ugly” IBM PC.