John Battelle's Searchblog

Hello,

My journey through the Internet, I came across your great site advised. Before I go back, I wish you a loving greeting leave here.
We would be on a visit / entry in the guestbook very happy!
Many greetings
The KOBV-Pöchlarn
Handicapped Association

Before I/we can discuss, I need a bit more clarification of what you and Steve Rubel mean.

You use the phrase "real-time, conversational search".

Does that mean that the search is in real-time? Aren't most searches in real-time? Or does it mean that the search is of real-time data, i.e. what you are finding is sorted by topical relevant combined with recently posted microblogs (tweets)?

And by "conversational search", do you mean that the search itself is conversational, in that you engage with the search engine in a back-and-forth, query-as-a-dialogue interactive information seeking process? Because that would really be exciting! Or do you instead mean that the search is a normal, 1.7-average-word search, just that the data-type being searched is "conversation" data?

And Steve Rubel calls it "Social Search". Is it really social? Are my searches influencing and interacting with your searches, socially, inside the algorithm itself? Because there is a whole community of people working on search that is truly social; the process itself by which the search algorithms operate take into account the social connections between people? Or are we all instead doing independent searches, but just searching data which was created, authored, in a social manner (conversation-data)?

Terms and definitions are flying around the blogosphere fast and fancy-free, and it's not clear to me what anyone means by all these phrases. There are many different ways of interpreting this, and I have yet to see a clear delineation of many of the issues. So I haven't been able to make up my mind if I am excited or not. Some of the definitions/interpretations excite me. Others bore me.

Keyword search is interesting but contextual search is so much more useful.

For me, yes, twitter is the real time web. But there is too much noise, just like the "common" web.

Some initiative try to get the semantic of tweets and tag them.

See http://tweetag.com

Do you know that we are reaching the 2 billionth tweet ? See http://www.tweespeed.com which try to estimate the speed and the final date ...

Hey Jon,

Unless things have changed dramatically over at Twitter, there's a question about the theory that search will become the nexus of the twitterverse:

Will it be able to?

I saw @ev speak at the Churchill Club in December, where he was surprisingly candid about how buggy the whole Twitter system was. When an audience member asked why it was so hard to find search.twitter.com, his answer was "if more people used it, it would crash our servers."

Six months and a bunch of funding have passed, so obviously that's a problem that's been worked on. Search is on the home page now, but is that because the scalability problem has been fixed?

I believe you stated here or on Twitter that "they've got to address the abandonment problem." They do. People sign up, don't get it, leave. Big problem.

Twitter knows they have to fix this too. It's my fear that they are pushing search to the fore to help fix the abandonment problem and not because the scalability problem has been fixed.

I, for one, do not live in a "tworld" that is free of the blue whale.

@adamkleinberg

@adamkleinberg

>>>Twitter knows they have to fix this too. It's my fear that they are pushing search to the fore to help fix the abandonment problem and not because the scalability problem has been fixed.

As far as I know, search.twitter is still running on a different cluster of servers from the main application, with an archived dataset. If Twitter can provide as much value from search as they do from their main application, they should be able to take load off their main application. Likewise if they can provide value beyond trending topics through groups and filters to unregistered users.

One of the important lessons of early Web 2.0 applications was to provide a useful function for unregistered users, and Twitter is getting there over a bumpy road. I'd be curious to see stats that go deeper than the Nielsen numbers - like do people get pulled back into twitter, do they get funnelled into 3rd party applications, etc.

Twitter users are undergoing large-scale change management, and Twitter's success or failure could very easily hinge on teaching new users how to use hash tags, reimplementing retweet or direct message. For the most part they're doing it right, it's just that for many people twitter is becoming an essential commodity. At least in this context, search already is an essential commodity.

"Twitter's success or failure could very easily hinge on teaching new users how to use hash tags"

Strongly disagree. Search should be good enough to work without them.

While early adopters may like the command line syntax of twitter with d @ # and RT, these are big steps back in usability.

Usability in the Web interface needs to become better. e.g. typedown against my following list when i send a direct message.