Comments on: Algorithmic puzzle

By: washburn

washburn — Mon, 15 Sep 2008 19:14:45 +0000

@Matthew: I haven’t had a chance to think through all of the code presented, but it does at least upon a naïve inspection seem to have the correct behavior and running time. So if nothing else it looks like a good start. Thanks!

By: Matthew

Matthew — Fri, 12 Sep 2008 22:39:43 +0000

This any use?

http://haskell.org/pipermail/haskell-cafe/2007-August/029816.html

http://twan.home.fmf.nl/blog/haskell/Knuth-Morris-Pratt-in-Haskell.details

By: washburn

washburn — Mon, 08 Sep 2008 18:55:42 +0000

@thom: No, we can't. The Seq type already has a length method, it is just not guaranteed to terminate. Furthermore, we can't use instanceOf to decide what to do because we can't know ahead of time all possible classes that might mix in the Seq trait.

By: thom

thom — Mon, 08 Sep 2008 18:46:03 +0000

i don’t know enough scala to give a proper answer. most java code that has to deal with a situation like this however would use instanceOf for example:

int getSize(Iterable iterable) {
if (iterable instanceof Collection) {
return ((Collection) iterable).size();
} else {
// loop…
}
}

can you distinguish between definitely finite and potentially infinite sequences in a similar way using instanceof in scala?

By: washburn

washburn — Mon, 08 Sep 2008 09:26:28 +0000

@Walter: Sorry, but someone else will have to do that. Frankly, I shouldn’t even really be working on this bug at all, as I have a lot of work on Scala Classic to do and the deadlines are already too close.

By: Walter Chang

Walter Chang — Mon, 08 Sep 2008 09:19:26 +0000

i dont have an answer for you but do have one suggestion, though. please provide overridden versions so that indexOf on finite seq can be fast and on infinite seq correct. thanks!