A.M. Kuchling comments on ISI's obituary for Keith Uncapher.

While still at RAND, Uncapher designed the first time-sharing computer system for mathematicians. He led the packet-switching technology project and the RAND Tablet Project, a computer system for recognizing hand printed characters using a tablet and stylus. Packet switching research and development led first to the military's Arpanet, and then to the Internet itself.

In addition to networking research, at ISI Uncapher started a cooperative to share the cost of prototyping new silicon chip designs that has been used by chipmakers for more than 50,000 prototypes.

The prototyping system mentioned in the second paragraph is MOSIS. Users submit the design they want to try out as a layout file, and MOSIS staff combine the layouts from several customers into a single large layout. A set of wafers then gets manufactured from the large layout and diced into individual chips that are sent back to the right people. Well-known chip architectures such as MIPS and SPARC were first prototyped using MOSIS, back when their designers were grad students. MOSIS was also a direct inspiration for the MEMS Exchange's model of a virtual fab.

New York Times Obit and Interesting People mention.

[a klog apart Obituaries a la Blog]

icann.Blog points out Dan Gillmor's great essay on removing structural choke points as strategic defense.

In a world where rationality prevailed, we'd launch a new kind of Manhattan Project to remove the energy and communications choke points. We'd actively discourage a software monoculture that leaves us so open to cyber-vandalism and corporate power hunger. We'd work harder to establish more competition for telecommunications, not let the industry consolidate to a tiny number of players.

This needs the momentum and moral authority of "monopoly busting" and "homeland security." How do we create political and economic action around this?

 [a klog apart public policy]