Read on Twitter
Listening to your episode on Richard Hamming this afternoon, when I was stacking firewood, @ericnormand. Some thoughts: >>> https://lispcast.com/one-mans-view-of-computer-science/
When I was at high school, fifty years ago, we were advised not to go into software because within five years (then), computers would be so intelligent that they'd program themselves. People's projections of the future of software are this: WRONG. >>>
When I went to university, the School of Computing taught Pascal & Fortran; I chucked it after a year because I felt I wasn't learning anything useful, and instead learned #Lisp in the philosophy department under @peter_mott, because we wanted to automate inference processes. >>>
(I also, bizarrely, learned BCPL and assembly language in the peace studies department, writing programs to assist in running war games for nuclear confrontation simulation – not that either of those have been of long term use to me.) >>>
What I think this illustrates is that teaching students familiarity with the language du jour is just mistaken; the languages those who majored in computing were taught in my day had ceased to be mainstream within a decade. >>>
The surface layer of computing is in continuous turmoil, with new languages, libraries, approaches continually emerging, becoming briefly wildly popular, and then fading slowly into obscurity. >>>
But underneath, the fundamentals remain much more durable. The techniques for approaching and decomposing problems, for diagnosing faults, for assessing which algorithms will scale and which will not (and why): these are the things which are important to teach. >>>
Most of all, what people who study the craft (I see it as a craft) of writing software need to study is how to learn; how to adopt new styles of thinking; how to assess new tools and techniques; how to recognise underlying equivalences; how to cut through hype. >>>
I'd also say that it's relatively easy to see where the centres of excellence – where you can sit in a room with twenty people and find that, twenty years later, half of them are legends – were a generation ago; much less easy to see where they are right now. >>>
You can name them. Bell Labs, Xerox PARC, Martlesham Heath, MIT AI Lab, Cambridge University Computer Lab, a few others. I'm sure there are equivalent centres of excellence active right now, but that they are much less visible. >>>
Anyway, thanks for yet another enjoyable and thought provoking episode.
