Return Styles: Pseud0ch, Terminal, Valhalla, NES, Geocities, Blue Moon.

Pages: 1-

Imperative > pure functional programming

Name: Anonymous 2010-05-23 3:58

Who would intentionally write code where you can't randomly access memory?  People who prefer O(n log n) over O(n) algorithms.

Name: Anonymous 2010-05-23 4:12

Well that all depends on the size of kf vs ki and the practical limit of n, now doesn't it?

Name: Anonymous 2010-05-23 4:30

It seems you have been brainwashed by Clojure morons. Even looking at the most pure functional languages, Haskell has mutable arrays via monads1 and Clean has mutable arrays via uniqueness typing2.

1 http://www.haskell.org/haskellwiki/Arrays .
2 http://clean.cs.ru.nl/CleanExtra/report20/chapter4/s44.html

Name: Anonymous 2010-05-23 4:46

Name: Anonymous 2010-05-23 4:56

Purely functional is theoretically beautiful, but it's not the most practical. A good mix of functional and imperative  with some facilities for abstracting syntax is most preferable for me.

Name: Anonymous 2010-05-23 5:13

>>4
Purely functional means exactly what it says: every procedure expresses a mathematical function. The property you must give up for O(1) array lookup and update is persistence.

Name: Anonymous 2010-05-23 5:17

Name: Anonymous 2010-05-23 6:00

>>6
Why do you imply your making a point with your italics when you have not made a point.

Name: Anonymous 2010-05-23 7:24

I like functional programming because it allows me to guarantee the functioning of pure functions for as long as I understand the underlying logic. I can then spend the rest of my time figuring out mistakes in the impure functions or learning about how to optimize the software from naive implementations to more efficient ones.

Name: Anonymous 2010-05-23 8:08

>>9
Yes, that's the advantage of functional programming, but OP isn't talking about functional programming, but the radical form of functional programming called purely functional programming which rejects the imperative paradigm in full.

Name: Anonymous 2010-05-23 13:40

>>6
No, the property you must give up for O(1) array update is persistence, which is almost always (but not always) necessary for pure functional behavior.

You can always have O(1) array lookup.

>>10
GBT/pr/

Name: Anonymous 2010-05-23 13:45

Wear gloves, don't lick the stamp.

Name: Anonymous 2010-05-23 13:46

mv >>12 ../1274633168/

Name: Anonymous 2010-05-23 15:50

>>12
lick my anus.

Name: Anonymous 2010-05-24 11:04

>>10
I see. That's kind of stupid as it would be impossible to interact with that kind of software. Nothing could possibly enter data into the software after interpretation and nothing could possibly receive the results of the calculation. Oh, I see now, IHBT, GJ >>1-sama.

Don't change these.
Name: Email:
Entire Thread Thread List