Re:A Badass Way to Connect Programs Together
by perishabledave on Tuesday January 26, 2016 @08:51PM
(#10977684)
I think he's referring to how Erlang uses the Actor Model. An Erlang program consists of (possibly) thousands of individual lightweight-processes. In this model a typical task may be spread out over many asynchronous processes. You can think of it as a program thats made up of thousands of micro-services. Instead of a task taking place over a sequential piece of code, it takes places over many individual processes that coordinate by communicating with each other.
Re:A Badass Way to Connect Programs Together
by Gracana on Tuesday January 26, 2016 @07:58PM
(#10977493)
Imperative programming. Lists of statements, vs Erlang's declarative programming, where you describe the program without specifying control flow.
Re:A Badass Way to Connect Programs Together
by white-flame on Tuesday January 26, 2016 @09:01PM
(#10977724)
Erlang is absolutely a fixed in-order control flow programming language, on message receipt. It is (generally) referentially transparent, but not declarative.
Re:A Badass Way to Connect Programs Together
by Gracana on Saturday January 30, 2016 @11:06AM
(#11002323)
Oh, I don't know much about Erlang so I thought for sure that's what he meant. What was that supposed to mean, then?
Re:A Badass Way to Connect Programs Together
by xrstf on Tuesday January 26, 2016 @07:59PM
(#10977500)
Given that he is contrasting to Erlang, maybe he's talking about imperative languages instead of (pure) functional ones?
Re:A Badass Way to Connect Programs Together
by dwc on Tuesday January 26, 2016 @10:41PM
(#10978051)
Erlang's pattern matching allows a great many things to be done in a declarative style. For some problems you end up with a lot of pattern stuff and not all that much "normal" code. Or you pattern match through what would ordinarily be a lot of flow control statements, for each case writing very simple code to handle that one thing.
Since Erlang isn't a declarative language in the larger sense, that's my guess what he meant.
Context here:
"I guess I’d underestimated the difficulty of implementing Anything-over-TCP in a sequential language. Just because it’s really really easy in Erlang doesn’t mean to say it’s easy in sequential languages."