Doing IRC right

A couple of weeks ago (wow, it's been a couple of weeks!), a new twitter-like service called started up. It's open source, and incorporates a lot of "duh" features that twitter doesn't yet have, so I'm a fan in general, but I think it's doomed to fail ultimately because it doesn't have the community that twitter has--yet, and I don't think that it will, ever.

With another new 140-character micro blogging service I had a series of conversations with bear about these kids of sites, the social niche that they occupy and the technology that's behind them.

I should start by saying that I really like twitter, and I think that the ultra-short form totally has a niche. I think the problem with identica, which it inherited from twitter, is that their social plan is out of sync with their technological plan. [1]

Twitter has, of course, been plagued by horrible scaling problems. I think initially twitter expected everyone to actually answer the framing question "What are you doing?" So that each individual's stream would be like a more or less isolated "microblog." And their software was developed in this direction.

And then "the community," happened, and it turns out that what people really want from their twitter-like services is not a place where they can record "what they're doing," but talk publicly with a group of friends.

Enter the "fail whale:" the problem with this usage model is that people check it much more regularly, they update more regularly, and with the site not really built for this kind of usage, there are a lot of error messages, lost messages, and the like. While it's taken twitter a while to sort this out--and they are getting better--the problem remains that twitters users and twitters initial designers have/had two very different sites in mind. And because of the problems that twitter's had (and because of it's great success) there have been a lot of sites that basically duplicates twitters functionality.

This is all fine and good except that the fact that twitter is the wrong thing to copy. My conclusion from the discussion was that, "the next twitter-like service to make it isn't going to get twitter right, for once and for all; but rather, the next twitter 'killer' will get IRC right, at last."

IRC is the original internet chat platform. It's not without it's problems, both social and technical. Basically you have to have a live connection to use IRC, there's no "offline." It also doesn't scale much more effectively than twitter, and there's no way to filter/organize the community on IRC except through really route tools[^socirc]. The really interesting thing is that people are more prone to using twitter like IRC, and less like blogger or live journal (which you have to imagine is what they thought that it would be like.) [^socirc]: The issue is that if a channel on IRC gets overridden with spam, or annoying people, or just volume, the only option is to start kicking people out, or to start a new channel, which can be sort of draconian.

So what would "a better IRC be?" I don't know, I'm just a guy, right, but here's what I'd like to see:

1. Integration with the web in a way that doesn't suck. There have always been webapplets for IRC, but they have always sucked. This can't continue. 2. You shouldn't have to be online to record a conversation. 3. Unlike twitter, the offline apps need to be as good if not better than the web site. 4. Arbitrary rooms, "moments" and streams need to be constructed on the fly by users. The randomness of twitter is something new, that--now that we know it's there--is something we want. 5. Everyone should be able to intuitively construct filtering mechanisms. 6. Chat "moments" and cork-board "moments" should both be possible to construct, but 7. No one should have to think about the infrastructure (this is a problem for both IRC and twitter.) 8. Identity needs to be managed coherently. Jabber/XMPP seems like an ideal tool for this project. 9. Also identitiy management and community organization seems like the "niche" for private enterprise to fill, rather than infrastructure, which I think can be decentralized. 10. Threaded conversations. It's a must.

Anyone have anything to add to this?

Onward and Upward!

[1]In a way, I suspect this qualifies as "writing a job for yourself," in to your analysis. My secret superpower, it seems is to look at what people are doing, and then talk to the engineers about the social realities. Any web 2.x people who want this, be in touch ;)
comments powered by Disqus