A week ago, John Gruber linked to Twitterific, a client for Twitter written by Iconfactory (of redesign pixel art animation fame; oh, and they do icons, and stuff – for Microsoft, etc.).

Twitter?

I had heard of twitter before, but never looked much into it. The concept is so extremely simple that it may not have much appeal to must people directly; as far as I can tell, you only benefit from it when you make up your mind what you want to accomplish with it. Twitter is so negligible right now that it doesn’t even have its own Wikipedia page; arguably, this is more a matter of being in its infancy rather than permanent irrelevance.

What you can do with Twitter is to describe what you are doing right now. In itself, that doesn’t sound very exciting or useful; and it’s not. The strength is actually in the many means of accessing and updating this information.

As you would expect, you can log in via your web browser and enter your status right there, and watch that of your friends, or everyone else (assuming they’ve set it to public). But there’s many different ways: your AIM or Jabber client, several Twitter-specific clients (such as the aforementioned Twitterrific) and even your cellphone. The latter makes for some interesting scenarios. For example, while in town with your cellphone, you can easily find out if any of your friends happen to be nearby, thanks to their Twitter-delivered status updates that you receive. But personally, I make completely different use of Twitter.

A few months ago, I moved away from a multiple-protocol IM client (the venerable Adium) and back to “official”, separate ones, especially because of the frequent failure of file transfers in Adium, but also for other reasons, such as lack of audio and video chat support (if your laptop has a built-in webcam, you might as well use it now and then, right?). Moreover, I had to use at least two apps anyway: Adium and Skype. Sadly, having separate apps comes with its own headaches; right now, and most of the time, I have iChat (AIM and Google Talk/Jabber), Microsoft Messenger (MSN/Windows Live), Yahoo! Messenger (Yahoo!) and Skype (well, Skype) running. All of them have their own UI that’s similar in some aspects and different in others (cmd-L, for instance, shows you the contact list in MS Messenger, but logs you in or out in iChat – an annoyance I run into all too often). All of them have their own log formats (and while Skype doesn’t fully index it in Spotlight, Yahoo! currently doesn’t log at all, though this will be remedied, including indexing, in a coming-soon release). And, to bring us back to Twitter? All of them have their own status message.

Fortunately, two of them, iChat and Skype (and therefore AIM, Google Talk/Jabber and Skype) have an AppleScript API to let you set the status message, and an easy-to-use one at that. And thus came my thought: what if I were to use Twitter as a framework not only for letting people keep up with my status on the web, on their cellphone, and wherever; what if I also use it to have one and the same status on my instant messengers?

This illustrates, I hope, how the simplicity of Twitter is not only deceptive, but also intentional. It makes for a wide range of usage cases. The more complexity Twitter were to add, the more specialized and constrained the result would be at risk of becoming.

That doesn’t mean they shouldn’t add features; they might very well have to. Indeed, the ultra-minimalist interface clearly has its limitations. Eric Meyer has described some today. Not only that; he also comes with another good point: because Twitter leaves so much up to the user, some users inevitably won’t agree with others on what constitutes appropriateness. Says Eric:

Then there’s the bonus reason I want to throw the whole thing into my bit-bucket: the way people are using Twitter right now, it’s rapidly becoming the most inefficient and unusable version of IRC ever. Look, people, if you want to chat, then get a chat room. You know?

I.e., people aren’t just being a little too “update-happy”; they’re also using the platform for something it clearly isn’t ideal for: an unusual kind of chatroom. This can be downright disruptive to others, especially because, as Eric points out, Twitter lacks any way to have you filter out what kinds of messages you wish to see, and which ones you’d rather skip. The whole concept of a “message kind” is far beyond the realm of Twitter; indeed, metadata, conceptually, is certainly an entirely foreign concept to this service. This is, ironically, just as much a flaw as it is a unique advantage. Whether Twitter’s admins will eventually step in and define boundaries, or whether the software and APIs will eventually see enhancements to add keywords or categories to your latest “tweet”, I don’t know. For now, we’ll have to deal with the funny situation where your “friendships” on Twitter are not defined by whether you like or strongly dislike a person, but whether you believe their updates are relevant enough for taking the time to wade through them.

In other words: the concept needs work. I can definitely see some very nice usage cases, and it might very well become one of my least dispensable tools, but even though it is among very few “Web 2.0″/”social” websites in that it doesn’t feature ever-abused “BETA” moniker on its logo, it happens to be one where, perhaps, calling it “a work in progress” would have been appropriate.