Thursday, October 2, 2008

(microblogging about) naked lunch

This post can be summed up in a picture.

My aim is to get across a perspective on microblogging that i think has some import. Let's do this bottom's up, i.e. start with an experience and then work upwards towards the cultural and conceptual context in which the experience has deeper significance.

The experience is 'mixing'. i noticed that the Twitter interface does not allow me to perform certain experiments with the 'mix' of feeds i follow.
  • i cannot change the font characteristics of specific feeds
  • i cannot change the interleaving characteristics of the overall mix
Some of the feeds i follow -- let's be frank -- have less information density. i still follow them because their overall quality adds something to the mix. Others have extremely high information density. Given this fact, i would like to rearrange the visual experience of the mix of feeds much like i might mix audio tracks.
  • i would like to turn up (resp. down) the 'volume' of certain tracks, by changing the font characteristics. Some tracks need to be big and bold some need to be red and italic others need to be small and grey .
  • likewise, i want to be able to place my tracks in 'sonic space', e.g. pan them left and right, which might be likened to interleaving characteristics of the feeds. For example, i might want to interleave Fred and George Weasley's feeds at the character level while i might want to interleave Ron and Hermione's feeds at the word level and Harry and Dumbledore's feeds at the 10 word level.
i submit that the user experience for creating and exploring different mixes could be nicely captured using a mixer metaphor. We might imagine that there is a display space -- a console of sorts -- where we see the mix -- think tag clouds. This is fed by an N-channel mixer (where N is small and manageable, like 16 or so). The mixer is connected to a bus -- the larger group of feeds the user is following. The user can select which feeds from the bus go into the mixer. The dials on a channel (which gets a feed) correspond to setting font and interleaving characteristics.

The point is that experiences like Twitter are really a step toward -- and beyond -- what Burroughs was talking about when he suggested "cut ups". The very fixed mixed we get from Twitter is still a view into the real-time stream-of-consciousness of a group. We don't really know how to get information out of the groupstream, in an efficient or even particularly effective manner. So, let's approach this like children and scientists: set up a lab and play. The mixer UI allows us to experiement with different ways of pulling information out of the mix, and creating different mixes.

The Twitter folks should take note: this sort of laboratory makes the experience considerably stickier. As soon as we add some element of control (mixing) to some element of chaos (the feed streams) we get creativity. That is incredibly addictive. Moreover, it starts to suggest new sorts of forms. There's the live DJ who creates astonishing mixes of live feeds, but there's also replay and feedback. And, there's catalogging and navigating the mix settings themselves.

The underlying point, though is that we are at a cusp where we are developing the technology by which the group and the individual can get a much better view of each other. They can experience each other as peers. We've been learning for the past century or more how much 'individuals' are really collectives (society of mind, organization of cells) and we've institutionalized a legal view of collectives (corporations) as individuals. But, we haven't yet found good ways for persons and collectives to regard each other as peers, though fruitful dialog.


leithaus said...

a friend commented via email -- I still do not grok "interleaving" by character. I'm a very literal guy (pun intended). That sounds like gibberish to me. How is this done so that the output is comprehensible?

leithaus said...

Here's my answer:

As if the word salad that is Twitter is comprehensible at the current level of interleaving (> 140 chars / entry). The point is to start looking at the mix like it's a picture, not like it's a piece of text.* Character frequency yields all kinds of information about a person and a crowd. Here's an example from the anthropology literature.

Which of these two (made up) words

* tikiti
* maluma

sounds round? Which sounds pointy or sharp?

It turns out that there is staggering agreement on the answers, across cultures. If the crowd is using a bunch of "t's" and "k's", it might just be like when a rabbit grinds its teeth or a horse flattens its ears or a dog raises its hackles.

* Or rather, that text and pictures were never so far apart as one might have imagined. Large text streams start to take on more analog characteristics. A guy who has done a really deep analysis of symbol systems is Nelson Goodman: The Languages of Art.

pseudonymous said...

more like tinny and woody...

:) I still don't see the value of mixing words, but _having_ that capability as a slider wouldn't take value away.