Thursday, May 20, 2004

For the last three years, my sole window into the digital world has been some kind of 15-inch Powerbook G4. Great machines, but the small screen was getting increasingly unwieldy with the many applications that I keep switching among. In my extended lunch hour today, I bought an Apple 20 inch Cinema display to more than double screen real estate in my office. Wow. Why didn't I do this a long time ago? 6:33:23 PM

The weather was very nice at lunch time today, and I didn't have any appointments for a change, so I indulged in a bit of random browsing at the delightful Penn Book Center. Found Window Seat: Reading the Landscape from the Air by Gregory Dicum. Great typography, images and layout on two of my favorite topics, geology and geography. Yummy. 6:25:53 PM

An unfair article about Linus Torvalds and his role in developing Linux. You can do a very quick barn raising in software, even of a substantial piece like a Unix clone. But Linux wasn't written by one person in a few months, it's been in development for a decade, by a group of developers. [Scripting News] What is most clueless about the article is its unthinking acceptance of a maximalist view of "intellectual property." Anybody who knows anything about the development of operating systems 1960-1990 knows that operating system designs, algorithms, and data structures circulated widely in the research community. Tanenbaum's MINIX was but one example of operating system code created for the purpose of teaching those ideas, and its success should be measured by how effectively it helped spread the ideas to other researchers and students. If Linus learned something from MINIX and put it in practice in early Linux, that's good news: reinventing the wheel is a terrible waste of human creativity. Copyrights, which concern the expression of ideas, have nothing to do with this. Patents, fortunately, did not intrude upon the scene but lightly until much later (the Bell Labs SUID bit patent being a notorious exception). And of course Unix itself was based on many operating system ideas invented and implemented by others. It may strike newcomers as surprising that most if not all of the operating system and utility programs used in the 70s and 80s were easily available in source form from the vendors for maintenance and adaptation purposes. Improved operating system and utility code circulated widely among users of particular machine types. It was not open source, but it still played an important role in the spread and advancement of operating system concepts and expertise. Even though I never worked on operating systems as a main activity, I was at different times involved in tweaking operating system code and utility for ICL 4100s, PDP-10s and PDP-11s. Much of what I know about operating systems and Unix in particular I learned that way. [Andrew Tanenbaum has now posted a note on the relationship between MINIX and Linux with an excellent first-hand account of how operating system ideas developed and spread.] 11:30:25 AM