Comments on: Is Google working on a way to search Twitter and other microblogs?

By: Amir Khan

Amir Khan — Thu, 17 Sep 2009 19:00:54 +0000

Hi Dean, Thanks for the post. Very useful, I referenced this article in my own post today:
http://apoint2share.com/?p=43

By: Chris

Chris — Mon, 15 Jun 2009 04:09:56 +0000

I agree with Howard H. Why not put some adsense on my twitter page, and give me a cut? I would post night and day, if I could make some money off of it.

Chris
http://worstiphoneapps.blogspot.com

By: Howard H

Howard H — Mon, 15 Jun 2009 00:42:59 +0000

It would be interesting having adsense/adwords on twitter.

By: Peter Antypas

Peter Antypas — Sun, 14 Jun 2009 22:18:52 +0000

Andrew is right. Several years ago, eBay asked Google to provide them with search, and Google failed. The reason was (as Andrew pointed) that Google's algorithm is optimized for tons of slowly changing content. They simply can't handle massive real-time streams. And, no, there's nothing spectacular about indexing News. There's actually very few news streams out there (Reuters, AP, etc). Compare that with potentially millions of tiny documents (tweets) per hour and you have enough to keep CS geeks busy for years.

By: D

Sun, 14 Jun 2009 14:34:27 +0000

Google already shows a lot of recent results for topics in the news using regular Google search. Google wouldn't have to throw out what it has to add on a section for tweets, just like it adds in a section for news results.

By: Andrew

Andrew — Sun, 14 Jun 2009 13:00:35 +0000

Not to put a damper on the notion, but Google's infrastructure technology is *theoretically* incapable of supporting real-time search. It is not even controversial, even among the other companies that use similar technology. Unless they have thrown everything out and started from scratch, I find it highly unlikely that any real-time search of microblogging is more than thin and superficial.

That said, I may be reading too much into this. Yeah, it might be possible for Google to have a search service that works adequately in a very limited scope with respect to microblogging. Nonetheless, the general problem of real-time search of various data streams is theoretically equivalent to another longstanding unsolved problem in computer science, indexing true spatial data, which also remains manifestly unsolved. I don't think many people fully grok the theoretical context of these kinds of applications nor the implications of solving the underlying technical problems.

To do this, Google would be required to discard their entire technology base as a theoretical prerequisite. I do not think many people would make that assertion.

By: stream

stream — Sun, 14 Jun 2009 11:59:52 +0000

i think no