I'm a developer in Melbourne, Australia, and co-founder of Hello Code.

Published Sun 20 April 2008

← Home

Twitter as social computer

Twitter is a funny little thing. It's one of those services that has been pared back to its most basic, essential ingredients — asynchronous chat/messaging. Or is it microblogging? Or lifestreaming? Whatever you'd like to call it, it provides the one service, and that's it. No file sharing. No 'social network' (well explicitly, anyway).

It actually sounded incredibly banal to me when I first heard of it. It seemed far too basic to be interesting, and once I did get into it, I found that it didn't really provide me with much. But as noted by Scoble, the service's value is in its users. Once I started adding people to follow, I started to get addicted. It gives you an insight into people that you don't get reading their blog posts, and an added benefit of finding out about news as it happens.

But the way in which Twitter lets people interact has provided other interesting developments as well.

Social searching

You might have heard the prediction that web3.0 is going to be about the semantic web — "understanding" and giving meaning to content, providing for a better search and better ways to relate data to people and to other data. This is all well and good, but the technology to add this layer of meaning isn't quite ready yet.

In the meantime, the best way of understanding a query and providing the right content is to ask a real human (not that ChaCha, who tried to provide human-assisted search as a service, got far with this). But still, the keepers of the semantic meaning are ourselves, humans. We are only trying to teach the system what we already grasp.

And this is where a very interesting property of Twitter comes into play. With an always-on pool of friends only a message away, suddenly Twitter is your own social computer. Tweet a question, say, "does anyone know which team won the footy tonight" or even "where should I stay in Sydney?", and if your social computer has "indexed" this data already, odds are you'll get a good answer. Try asking Google one of those questions and you won't get anywhere near what you wanted.

Now obviously the results aren't as formal as if you had queried a search engine, went to a news site or browsed a site which recommended hotels. They might not be quite what you are after either. But the humans on the other end can understand your question far more accurately than a search engine can, and with an added benefit of trust and authority — a recommendation from a real person carries far more weight, which is why word-of-mouth is such a valuable marketing tool.

Bringing back the command line?

Another interesting feature is the use of hashtags and @replies to give control and meaning through what is essentially a command-line interface. Replying to someone means putting an @ symbol and their username at the start of your tweet — a modifier built right into the command line, with no other way to do it. Putting a hash before a word (usually a noun) makes it into a tag that can be referenced across tweets, giving it some semantic meaning. These are two ways that the textual data provides control as well as content, which is the main feature of a command-line system.

But when you also add twitterbots to the mix, you start to see the sort of power Twitter's basic input/output system has. If you tweet @weather Melbourne, you'll get a weather report — just like calling a command line executable and passing it an argument. Twitter's system is ideal for this — as a basic IO system, it takes in a small set of data, and provides a response. That sounds limited, but it has so many varied functions that have only been half-realised yet. Of course, there is always the command line, Google, or some other GUI which can provide you with the same data, but I'm sure we'll see a benefit to having this data included as part of our attention stream.

An interesting ecosystem

It's funny what you can do with such a simple system. I guess when you create an open API that allows people to build bots, clients, and other ways to analyse the data, there's actually a lot to it. And I think this is only the start of the sort of clever social systems and lifestreaming services we're going to see (forget basic social networks, they're so last year). But maybe whoever said it was right — web3.0 really will just be about cutting out all the noise, so that we can hear the signal again.

To top