9 Mar
2025
9 Mar
'25
4:14 a.m.
On Mar 8, 2025, at 6:05 PM, Rob Pike <robpike(a)gmail.com> wrote:
I was at PARC in 1984, working with Dan Ingalls. I mentioned I was surprised that
Smalltalk had no concurrency†, that the UI (let alone the system) was completely
single-threaded. Only the window with focus could execute any code. Dan being Dan, he
immediately got to work making a form of concurrency happen, followed by a delightful orgy
of researches playing with the new toy. I loved it.
Because: sometimes in isolation you miss important things going on in the outside world.
Surely they must've read papers on concurrency & were aware of CSP, monitors, the
Actor model etc?
A few years ago at a dinner I had asked Don Knuth whether he was going to write any books
on parallel algorithms. Alas, I don't recall his exact answer but he didn't seem
keen on the idea -- I was a bit surprised but thinking more about it, it made sense.
[Still would like to see someone attempt a Knuth style encyclopedic treatment to the
subject of concurrent/parallel algorithms!]