Information Technology

(The following is a kind of current-state analysis, maybe leading to my long intended anti-WWW rant. It might become more interesting in a few years.)

Honestly, I think Twitter is a broken technology. 140 characters are way too few to provide useful context. And as URLs eat from this character pool, the urge for link shorteners simply leaves this service behind broken, IMHO. Just look at this plot of the distribution of tweet lenghts:

I think this plot is not funny, it’s serious.

Well, some misuse Twitter as a chat, although IRC would be the technology of choice for that. I mainly use it as a news stream. In fact, Twitter changed their orientation from individual “Took a dump, ate a banana” status updates to personalized news. Although RSS or Atom are dedicated technology for collecting headlines and articles (and keeping their read/unread state), Twitter provides a unified time flow (that passes by, no matter if you read it or not). And while RSS/Atom is a kind of pull technology (i.e. you have to look at different feeds and articles for yourself), Twitter is a push variant where all elements meld into a single data stream.

The problem is, in engineering terms, there’s so much noise in the data; there’s so much irrelevancy occurring in the timeline, at least in mine. I followed more and more people, like “Hey, this guy developped that app, and there’s a link to his Twitter”, or I chose to follow various companies when I noticed they had an account at Twitter. But more and more often I asked myself where a certain guy/gal I followed actually came from.

I’m wasting too much time at Twitter. Meanwhile, I’m following more than 250 people. (Yeah, I think those who follow >300 are simply nuts. How the hell do they handle their timeline? How can they claim Twitter is not a distraction?)

Me wants:

• More than 140 characters please. 200 or so should be OK. Or, better, do it like Google Buzz or Facebook and just collapse longer posts.
• Attach personal notes to provide the context about who this Twitter user is or why it was interesting to follow him/her. Their own bios are often useless.
• Some sort of global maximum tweet rate to allow skipping of irrelevant tweets
• Give weights to followers: “Important” followers should always be in the timeline, while tweets from others are allowed to be skipped if there are too many tweets; also, some followers might be muted, but not unfollowed (i.e. some kind of bookmarking).
• There should also be a rate per user. If one suddenly starts to reproduce audio content from e.g. a conference talk, I shouldn’t be bugged, please.
• The decision about relevance should be made similar to the My6Sense client. This innovative client aggregates from one’s Twitter, Facebook, Google Buzz and Google Reader subscriptions and automatically learns the topics that are most interesting for the individual.
• This automatic learning should be done using the read messages/articles (like My6Sense does) and/or by buttons in a “plus”/“minus” manner to prevent “accidentally” read articles from becoming relevant.
• URLs should not eat from the character pool and be placed in some kind of footnotes. Or, rather, make them the usual linked words.

Of course, a lot of this already sounds like RSS. But there are tweets that are status updates per se and don’t contain URLs.

It appears to me that Twitter and RSS (and maybe the whole social *BUZZWORD ALERT* media) are becoming “the HTTP of a new WWW”: It’s up to new and upcoming user interfaces to aggregate, weight, filter, sort and manage content coming from various data streams; new user interfaces for new devices, but also for the “old scholars”. I noticed recently that things are becoming better for me when using different Twitter—or rather, aggregator—clients, like TweetDeck or My6Sense. It’s becoming important for me to not just scream or hear screams, but to consume and provide relevant information.