Header Ads

Search Engine Galore

Yes, Bing, Microsoft’s revamped search engine, has launched to much hype. We got all sorts of reports from early adopters ranging from “meh” to “it’s okay” to “it’s good and I’m making it my default search engine”. While we’re all for change, should we really be switching to something sub-optimal just for the sake of change? Probably not. But is Bing sub-optimal? Let’s look at what the search engine offers. Like Google, it’s fast, it has image search, news search integration – which for local news is better than Google News – and your regular maps, videos and products search. It even offers auto suggestions as you type your query. However, in spite of all the glitter, a search engine is only as good as the results it’s able to deliver, right?

To be able to look at this objectively we need to first analyze what the various search scenarios for most of us normally are. At times we’re looking for some yes or no (factual) type of information. For instance, if you were to type TechQuark, you’re probably expecting a search engine to take you to our web site. There is nothing better or worse that a search engine can provide other than a link to the TechQuark web site and maybe a Wikipedia entry. But what about a search where you don’t really know what you’re looking for. In this case, your knowledge of a particular subject will be restricted to the search engine results. This brings in the question of trust. Here is probably where Bing stands to loose out to Google. When it comes to such searches, users will constantly find themselves cross checking with the search giant to see if the older more experienced search engine gives them better results. There is always the underlying doubt – “Am I missing out on something. Will Google be able to show me something else?”

Let’s put it through a small test. First to level the playing field, we change the location of Bing to United States. This is because the search engine has become notorious for geography based content filtering. There is no way to turn the censor off except to change your default location. Bing definitely looses a few points here. Next we threw a set of search terms at both the contenders. The terms include a deliberately misspelled word (yes many folks use search engines as spell checkers), a single answer query like “Population of India”, and a more research oriented search term such as “server virtualization”. In the first query both search engines perform equally well. So we kept entering many more misspelled words. Google proved better at its “did you mean?” suggestions. For “Population of India” the results were quite curious. Google returned 1,147,995,904 while Bing showed 1,147,995,900 – from different sources! Next comes the more open ended test. Both performed quite well, returning almost identical results. Except Bing showed three sponsored results eating up valuable screen space. Google had a few news results too, which Bing did not in the main results page. Bing, however, has a nice little feature that lets you expand the snippet text by moving the mouse over to the side of the search result.

This little back of the envelope test is of course not perfect but it is indicative. Even if Bing is just as good as the currently reigning search champion, will we trust it enough to be exhaustive and accurate? Most likely, we’ll keep benchmarking it against Google for days to come. Are the breathtaking backdrops with tidbits of interesting click-able information enough to make it the search engine of choice? Some folks might switch, enamored by the novelty and buzz. But most will keep coming back to Google at least for now. Until of course Google does some major blunder down the line.


Wolfram Alpha is not a search engine. It’s a computational knowledge engine. Confused? Well, in a nutshell it is designed to simply give you answers to factual questions. For example if you were to ask it – the distance between Mumbai and Mangalore. It will not only tell you the distance as the crow flies, but also by road, the time it will take to reach there by regular aircraft, at light speed or even at the speed of sound. You won’t even have to enter the Mangaluru form of the name. It already know it! Sounds like the sci-fi computers we’ve read in so many novels right? It is exactly that – the character Data from Star Trek but not in the android form. So, in effect what Google or other search engines try to do with questions like Population of India, Wolfram will do much better. Conventional search engines will either parse language on a page and return many results, or if human intervention by way of tagging is involved, it will return information based on tagged data, or as a last alternative point you to web pages that should contain the answer. On the Wolfram web site it says its aim is to make “all systematic knowledge immediately computable by anyone”. Wolfram alpha has aligned all the data it has available to it in such a way that it can actually make sense of it; compute it. It has built tiny models of computational knowledge to compute human fields of data systematically. So you can ask it to tell you anything from “Number of internet users in Asia” or how much time it will take you to type 3000 words at standard typing speed, or find out which year in history was the name John most popular and it will actually compute the answer; replete with charts, graphs and any other information you might require. Of course by now you might have guessed that it is not pre-programmed with all sets of questions and answers that can ever be asked – that’s just not feasible.

Whatever little gimmicks we mentioned so far are simply wolfram being told to be a cheap conjurer of tricks, and for the world to marvel at. There is however, immense power and value in this technology. It can be used to mash up and find co-relation between stock movements, create charts and graphs to make sense of tons of socio-economic data and reveal patterns. The domains in which the engine delves into are vast too. You have engineering knowledge computation, music – ask it to play any note, finance, geography, even the human genome sequence. These are just some of the applications it can have. There are plans to leverage it to create custom versions for internal data, connections with other forms of content, and deployment on emerging mobile and other platforms. As it’s database grows and computational knowledge increases it will only get smarter. As of now it has 10+ trillion of pieces of data, 50,000+ types of algorithms and models, and linguistic capabilities for 1000+ domains.

So if it knows so much will it take over the world? No, because it doesn’t have a consciousness. But if everything is answered so easily it might make us brainless retards. Google only helped us locate stuff; we still had to make sense of it for ourselves. This little puppy on the other hand is going to answer everything for us.

Google Wave

Google wave is a cool piece of technology that centers around live and concurrent editing and workflow management. What this translates into is a personal collaboration and communication tool that is fun, and perhaps an extremely powerful a platform that can be used in many different ways. Now, there is an hour and a half long presentation available on the Google wave web site. But without going through it for you to get an idea of what wave is lets put forth this analogy. Think of a blank canvas. You and a bunch of other folks are simultaneously working on it. But you’re not just using paint and brushes. Instead, you have a multitude of tools at your disposal such as videos, pictures, maps and so on that you can pin up to the canvas to better present your point. And all this happens in real time. Say one of your fellow ‘artists’ leaves. He or she can come back and rewind and playback the flow of work to see what was missed. Okay now lets get down to the real world. Your canvas is actually based in HTML5 and can be run in several of today’s browsers. To speed up communication and give added functionality it features a spell checker that auto corrects on the fly. The spell checker draws on a huge contextual linguistic model, so it knows the context of words being used and corrects them appropriately. The interface consists of panels which they call – you guessed it – waves. These waves contain conversations that can be edited in real time! Forget about hand picking text to first quote and then reply. You can just pin up your response to any part of the original communication. What you’re typing gets bounced off a server straight to the other recipients of participants of the wave. If you wish to add new participants to a wave its as simple as dragging their picture from a contacts panel on the left.

All this free for all communication might just not go down well with those who are a little reticent. That’s one of the first things that comes to mind when you watch the wave presentation; “hey I don’t want people to see what I’m typing while I’m composing a mail”. Understandable. And the developers of Wave are not unmindful of that. They’ve incorporated a feature that hides you from the rest of the Wave. Perhaps the best thing about wave is that it can be embedded on other sites. Another interesting feature of wave is that discussion and content collaboration can be done on documents in the same tool i.e. wave. So you don’t have to follow the traditional Wiki and document format.

Wave is open source. That means that developers will get their hands on a set of external APIs with which they can build widgets for wave (which they call extensions) and even discover better ways of using its functionality. App developers only have to worry about coding XML on the client side and the server takes care of updating over the wire. Collaborative games are also on the drawing board amongst other things. There is also the scope of developing server side extensions using bots. Google demoed one such bot “Rosy” that can translate 40 languages on the fly! Since Google is keeping the much of the code open source, other corporations are free to come up with their own deployments of the wave system. By the time Wave formally launches, in a year, it might very well become a refined and extremely powerful tool. We’re eagerly waiting.

No comments: