M's

mathematica, micr*formats, and the next unix

[Attention conservation notice: Some incoherent ideas about using graphical instead of textual data for human-machine and machine-machine interaction in an OS.]

The outputs of many Unix commands follow the Microformats principle: Designed for humans first and machines second. Humans is a bit debatable in the Unix context, but still, the data is useful to the geeks among us, and also to scripts that process it.

ms@zulu:~/mov$ ls -lh
total 9.4G
-rw-r--r-- 1 ms ms  1.4G 2007-05-21 16:21 Akira.avi
drwxr-xr-x 2 ms ms  4.0K 2007-05-26 10:48 Amarcord
drwxr-xr-x 2 ms ms  4.0K 2007-05-02 11:24 Ascenseur pour l'Echafaud
drwxr-xr-x 2 ms ms  4.0K 2007-05-12 03:55 As Tears Go By
drwxr-xr-x 2 ms ms  4.0K 2007-05-09 00:21 Blood Simple
drwxr-xr-x 2 ms ms  4.0K 2007-05-11 02:44 Fallen Angels
drwxr-xr-x 2 ms ms  4.0K 2007-04-30 15:23 Fishing with John
-rw-r--r-- 1 ms ms  701M 2007-05-20 16:19 Happy Together.avi
drwxr-xr-x 2 ms ms  4.0K 2007-05-16 23:38 Kukushka
drwxr-xr-x 2 ms ms  4.0K 2007-05-21 15:53 La Notte
drwxr-xr-x 2 ms ms  4.0K 2007-05-13 13:39 L'Avventura
drwxr-xr-x 2 ms ms  4.0K 2007-05-23 12:16 L'Eclisse

This is micr*formatted data. (We have to say micr*formatted, or Tantek will sue us.) 

The Mathematica folks are doing some seriously cool work extending a shell-like environment with graphics, and the the demos are amazing.

Can we make it so (and would it be useful) that the returned graphics are micr*formatted data that can be passed to another application?

think &

the little things

Sometimes a word or two are enough for a thorough aha!-experience.

One is the connection between (software) architecture and strategy researched by Carliss Baldwyn at Harvard Business School. Related to the research on user-driven innovation, they note that modular software architectures, by giving users more options, lead to more outside participation, and thus more innovation.

Michael Tiemann (VP and Grand Unification Theorist at RedHat), whose blog pointed me to this research, notes:

Can these effects be accomplished without open source software? Yes, to some extent, but why do things the hard way? Are there other things besides innovation that determine the success or failure of a high-tech company? Of course--management competence being one big one. But in my book, efficient, high-quality, long-term innovation is the fundamental competitive challenge in building a sustainable technology business, and thus the profit-maximizing strategy is one that encourages innovation, rather than extinguishing it.

Another one is Rohit Khare's idea to model failures as price shocks. In his presentation, he gives the example where the failure of a RAM chip leads to a higher "price" for storing things, which could incentivize programs to use different ways of achieving their goals, e.g. recomputing data each time instead of storing it in RAM. The basic idea is to arrive at an "invisible hand" architecture that doesn't require pre-planning for all possible contingencies.

cheap, distributed corporate identity

I used to be fascinated with Sony's corporate identity. They have managed to create a unique style over an extremely wide array of products and services (and sub-companies).

Google has their plastic logo colors, pastel background colors, and these balls...

You can basically create a Google-branded page by using some pastel colors and slapping these balls onto it. This is cheap and easily distributed corporate identity.