Hacker News Reader

Comment by bmitc | Hacker News Reader

Your only critique is that SICP is old. Computing power and performance are irrelevant to SICP.

Not only that. Now precisely because multicore is the only path to follow when we reach the end of the Moore’s law, we need more than ever functional programming. Is no coincidence that there are so much new languages that are trying to use those concepts, including a big revival of Lisp.

Functional programming doesn’t help with multicore. The real problem with multicore programming is data dependency, of which mutability is merely a special case. functional programming doesn’t help prevent you from depending on a certain computation and forming the bottleneck of your program.

There's a big revival of lisp?

Many ideas that come from lisp, and for which lisp was long ridiculed, are now mainstream, like garbage collection, closures, a big runtime, a huge library (nowadays lisp is criticised for its small library), typed data (instead of typed storage), exploratory development in an interpreter coupled with an optional compiler...

I don't consider languages with these features lisps, but I think the GP's statement is only quite mildly hyperbolic.

That's not a revival of lisp itself, just it's ideas becoming pulled into more mainstream languages. I wouldn't call Python having a REPL to be a lisp revival, but you can disagree.

Erlang and especially Elixir with macros are not too far off of Lisp and are two of the more popular functional languages.

Multicore is a continuation of Moore's law not its end.

This so much.

Reply on news.ycombinator.com