Twitter’s founders initially formulated it as a broadcast tool to publish announcements to your friends and to the world, and to that extent it works fine. But once dialogue is taking place, Twitter becomes a cocktail party where everyone has a megaphone. Twitter lets you shout in public—so imagine everyone trying to shout conversations with one another in a public space….on Twitter, anyone who might take offense is likely to overhear….its design stresses conflict and impedes consensus. Once five or 10 people who disagree with you descend, it is very difficult to keep the original conversation going, since Twitter’s threading of conversations is nightmarish. Not only do you lose tweet space as a conversation gains more members, but as tweets branch off with different combinations of people, it can be impossible for any one person to see the entire conversation.
Auerbach then goes on to provide a clear example of how, for online discussion, the venerable Internet Relay Channel remains a vastly superior space and notes:
IRC…offers everything for conversation that Twitter doesn’t: topic-based chat rooms that you can drop in and out of, a real-time roster of participants, and a single complete stream of conversation…..On Twitter, the line between discussion and harassment is slippery. As soon as people overhear something they don’t like, they can drop in. If a lot of people hear something they don’t like, you will get swamped, and since you are always alone on Twitter…it can be discomfiting if even one of those people is less than friendly. Moreover, even if your critics have the most impeccable manners, it’s easy to become defensive and even scared if a dozen people simultaneously and independently disagree with you.
Twitter is a verbal minefield that encourages harassment while discouraging productive conversation.
More than two decades ago, when I first began studying online discourse in the context of computer mediated conferencing systems (in particular, EIES), I took a class in user interface design with my then academic adviser, Murray Turoff. One of the central commandments of interface design for the systems we worked on was quite simple: the interface should roughly map on to the cognitive style of the user. That is, if you are used to communicating or problem solving in a particular way, one which ‘works for you’, which enables you to achieve your substantive aims, then the online system’s interface should approximate it as closely as possible. You should not have to change the way you communicate or solve problems in order to accommodate your system’s idiosyncrasies; it should be the other way around. Otherwise the system is guilty of the sin of task modification. (Remember the drunk searching for his car keys under the lamppost rather than the parking lot where he dropped them, because ‘the light’s better here’?)
A classic example: old school command line mailers used to ask you to enter a recipient’s address in the cc: field before you began composing an email, and then moved you on to the actual writing of the text. Once you were done writing, you could not change the address fields to add more recipients list – whether in the to: or in the cc: field. Now, we do not write and communicate like this; very often, we will think of who should receive our communiques once we have written them. Its content–dynamically generated by us–can lead us to reconsider who should receive it: ‘This might be of interest to X too.’ The email clients that soon replaced the command line mailers added this functionality – you can add recipients to your message before, after, and during the composition of a message. They better approximate and accommodate their users’ styles.
Twitter is guilty of task modification. It forces us to change the way we communicate–even if we are in public–simply to accommodate its design features. Auerbach is right: Twitter is broken.
Note: Of course, individuals have vastly differing styles of communication and work so an ideal interface would be tailorable by each user. More on that later.