Imagine there is a big ticket event happening right now in another part of the world. Unfortunately you cannot be there but you really want to find out what’s going on there. You could wait till the reports are published after the event, but wouldn’t it be better if you could find out right now what is happening? Or maybe you are researching an article or a blog post. Would it not help to know what people are interested in right now rather than what interested them in the past? Welcome to the world of real-time search.
What is real-time search?
The Web today is growing at a rapid pace. And so is the volume of information. Digest this: 35 hours worth of video is uploaded on YouTube every minute; thousands of tweets are published every minute. Web search helps us find information on any topic that is of interest to us. With a caveat. The results are a snapshot of a time in the past even though sophisticated Web crawlers can index information fairly quickly—so it is still not what can be called “real time”.
Here and now: How do you find out what’s happening?
Real-time search essentially enables us to discover information that is happening right now. And this should not be applicable to blog posts or articles that went live now. It is about someone sending an update from that event, or posting images right away to Flickr.
Generic search for the most part did not have the ability to provide us with this information. But the popularity of social and information networks has changed all this. Updates on Twitter and Facebook are in real time. Both Google (http://google.com/realtime) and Bing (http://bing.com/social) now have access to this data and allow you to search through it. They do not really rank the results, however. Both provide a live updating stream of results related to your search results.
Twitter search results too provide a live updating stream related to your search term. But it is plagued with the same problems as Google and Bing. Spam. Primarily because the ranking criterion is the publish time of tweets. But we can expect this to change in the near future. With Twitter analytics rolling out soon, we can expect some authority-based ranking. And since Twitter has the most complete access to its data, it would be wrong to dismiss it despite the shortcomings.
Topsy (www.topsy.com) is one service that helps us get past the failures of Twitter search. Topsy indexes and ranks results based on identified influential conversations (as determined by their algorithm). This goes a long way in separating the noise from the results. Since Topsy indexes all tweets and archives them, it is also most likely to find information that is not necessarily new. You can also filter the results on multiple criteria, such as latest results, popularity in the past hour, day, week, month, and all time. You can also choose to see updates from experts as determined by their algorithm.
Besides Twitter updates, Collecta (http://collecta.com/) also monitors the update streams of popular sites such as Wordpress, YouTube, Flickr and TwitPic. Results can be filtered by stories, comments on blog posts, updates or media type. This makes Collecta unique among the lot in that it can search through media updates as well.
The big difference with Facebook is that you need to be logged in to search. And then too there is no easy way to search just through the status updates. Once you search for the term of your interest you need to filter the results by “Posts by Friends” or “Posts by Everyone” that appear on the left-hand pane on the results page. Of course, the updates you see depend on the privacy setting of the person posting the update. But Facebook does allow you to concentrate on results from within your social graph. Something that is missing from the other services.
Real-time search is still a developing field that depends primarily on only one signal—publish time—to rank results. But as the area evolves and signals such as author influence and content relevance are included in the ranking algorithm, it will become an even more intriguing and useful way to search for information.
Write to us at firstname.lastname@example.org