3 ms·

I've been pushing functional collections for years (with implementations for Common Lisp and Java: http://www.ergy.com/FSet.html http://www.ergy.com/FSet.html)

I've been pushing functional collections for years (with implementations for Common Lisp and Java: http://www.ergy.com/FSet.html http://www.ergy.com/FSet.html) and I have to say this post amuses me greatly. For most of this time, I've had to deal with people being skeptical about them because they were afraid they would be too slow -- and now the word is spreading that they're actually faster! I think that's probably wrong. Well-implemented mutable collections are faster than functional collections, if for no other reason than that they allocate less. As others here have said, the argument for functional collections has almost always been on style and clarity grounds. In some cases there's a stronger argument because you need to have many slightly different versions of a collection existing in memory simultaneously; I have written a piece of code like that, but such situations are pretty rare. I don't want to sound like I'm laughing at your argument, which is all true. It's just funny to see someone putting it that way.

jlongster 11y ago
The posts here are missing context: I'm talking about immutable data structures specifically in the context of UIs. It's common knowledge in the React world that there is quite a performance improvement using them, because as you change data you no longer have to do expensive checks to figure out exactly what has changed. You just rerender from the top-down and only update specifically what has changed (which just requires pointer checks now). See http://swannodette.github.io/2013/12/17/the-future-of-javascript-mvcs/ http://swannodette.github.io/2013/12/17/the-future-of-javasc.... In general, sure, if you can't take advantage of these semantics then there is going to be overhead and you just gain simplicity. But in certain domains we can take advantage of this simplicity for big performance wins.

jlongster 11y ago

The posts here are missing context: I'm talking about immutable data structures specifically in the context of UIs. It's common knowledge in the React world that there is quite a performance improvement using them, because as you change data you no longer have to do expensive checks to figure out exactly what has changed. You just rerender from the top-down and only update specifically what has changed (which just requires pointer checks now). See http://swannodette.github.io/2013/12/17/the-future-of-javascript-mvcs/ http://swannodette.github.io/2013/12/17/the-future-of-javasc.... In general, sure, if you can't take advantage of these semantics then there is going to be overhead and you just gain simplicity. But in certain domains we can take advantage of this simplicity for big performance wins.