It's No Joke: Real Time Search is a Big Deal. | Hacker News Reader

It's No Joke: Real Time Search is a Big Deal.

intridea.com

19 pointsmbleigh10 comments

How much will it take for Google to actually compete in this area, specifically against twitter and their summize search? Even if they didn't get access to the firehose, this data is already available via RSS feeds, Google would just need to crawl the RSS feeds of twitter users, which are (currently) most likely much fewer in number than the number of web pages that google crawls per hour, and update a temporal index (because as tweets age, they cease to remain "real time"), and show these along side their actual search results. "Your search for 'super bowl' returned this wikipedia article, espn.com and these currently active conversations". I suggest RSS feeds here because they are in the right format for extracting metadata about real time content (datetime, author, text).

In fact, if Google (or another major search engine, but right now only Google could do it) did change their interface to include "real time" results, it would encourage the rise and spread of more twitter-like services (because real time content would be increasingly easier to find). In fact, if Google included temporal RSS content next to the, for lack of a better term, archival content, and indexed comment threads also, the real time web would transcend just being provided by twitter.

Come to think of it, didn't Technorati attempt to do this with blog content, provide a "conversational pulse of the internet"?

Right now, Google classifies all content they receive, no matter when they receive it, as archival, and its freshness is only used to increase the perceived relevancy to float younger entries to the top.

The problem is that real-time search is a fundamentally different proposition from archival search. In real-time search what happened in the last 5 minutes can be more important than everything that has happened before.

Google could certainly add real-time search by gaining access to the firehose (and that would be pretty cool), but simply running against the RSS feed wouldn't provide the kind of real-time updates necessary to keep up with Twitter search.

I think google could fake it enough with twitter RSS feeds and by aggregating RSS content from other souces, call it "real time" and provide an integrated platform that piggybacks on what people already use. It doesn't matter that it can't keep up with the "real time" that twitter provides, people would clammer to be part of google's real time offering.

So, O'Reilly had a glitch, everybody and their dog twittered about it, the author used twitter search to get the info early.

I don't see the Big Deal here anyhow. How about just shooting off an E-Mail to O'Reilly and ask?

What if something happens where there isn't an omniscient authority? Or if the authority is unresponsive?

The cool thing (according to the author) was using twitter search to find too many submissions accepted for RailConf...

Well, can you think of an example, even a contrived one? I can't anyhow.

He used twitter search to find everybody and their dog twittering about it. I think his argument is that google search is not set up to provide this kind of information as quickly.

I think the idea is that it's easy to find a pattern when you are looking for one, but in this case, the author wasn't really looking for a pattern. He typed some stuff into twitter, and the pattern emerged in real time.

That is pretty cool.

(It would be even cooler if the author didn't bold every other word.)

"we built an API. 2000 pieces of software are capable of sending updates. Summize was the most important one" @ev #ted

http://loiclemeur.com/english/2009/02/evan-williams-of-twitt...

Reply on news.ycombinator.com