/ api

The Thing About Twitter

The thing that bothers me most about Twitter’s API announcement is that very few of the most useful features of Twitter were actually their ideas.

  • Hashtags.
  • Retweets.
  • Location.
  • Search.
  • Lists.
  • Conversation view.
  • Inline images and links.
  • Short URLs.
  • The entire mobile user experience.

All of these appeared in user behavior or other clients or other apps first, and Twitter picked them up into their clients, or bought the client outright.

I’m for picking up features from clients. I think it’s fantastic: a feature can be effectively tested among a small group via a niche client and if it’s compelling and popular, well, you may want to steal it for the whole userbase—or not, of course.

The comparison that sprang to mind was the Firefox add-ons ecosystem. Something can start as an add-on (say, Tab Candy or Personas) and if it’s popular and a compelling feature, sometimes it gets picked up into the product (Panorama, Personas). Other times (Firebug, Adblock Plus) it is incredibly compelling but not for everyone, and it doesn’t. (Or, in the case of Firebug, some of the ideas, if not the implementation, were picked up.)

The point of all that, to borrow a phrase from Jeff Chausse, is that thinking you will “inevitably make better apps [is] epic hubris.”

If 90% of user access to Twitter is through official clients, like they claim, then they are overstating the threat of “fragmentation” of the user experience, and doing so in a way which discourages anyone else from trying to do better, or simply different. (Unless, of course, they aren’t putting any technical teeth behind it, but in that case why say anything at all?)

Considering how much of the current Twitter experience was invented outside the company, it seems like this is indeed epic hubris.

NB: As all content on this blog, this post is my personal opinion and should in no way be construed as representing my employer.

The Thing About Twitter
Share this