I'm coming to the conclusion that CS papers need to be printed.
They don't suit ereaders (PDFs don't reflow), they're too small on smartphones and I find a laptop less portable.
I suppose the clue is in the name!
miniblog.
Related Posts
One interesting property of both stdin/stdout based REPLs and RPC based REPLs is that they need to support asynchronous events.
In both these programs, I don't need to wait until the function is done to see the output printed. It's not sufficient to read-eval-wait-print-loop.
I have a theory: Suppose I created a language Foo that printed execution time at termination:
$ ./hello
Hello World!
(Finished in 0.6 seconds)
I'm sure that Foo developers would be particularly sensitive to performance (for better and worse).
When I started out writing lisp, I found the distinction between foo and 'foo tricky to grasp. It bothered me that 'foo wasn't printed 'foo.
I think this is easier to learn when symbols are printed differently. If 'foo is printed user::foo it's easier to grasp what a symbol is.

