[Jprogramming] Code review / Euler

[Lau B. Jensen]

Dear J-Community,

It has been remarked that solving Euler problems using Clojure is almost 
cheating and I
followed up by saying that solving them using J felt like I was being 
cheated :)

I stepped back, started reading 'Learning J' and retried. This time the 
solutions come quicker
but feel...less than idiomatic. I'm hoping a few friendly J'ers will be 
able to spot where Im going
wrong here.

#1: Find the sum of all multiples of 3 and 5 below 1000

 +/~. (((0=3&|) # ]) i. 1000),(((0=5&|) # ]) i. 1000)

What I really dislike is that I'm repeating myself throughout, I want

3 5 (some func) i.1000

But how would I go about that?

#2: Sum of all fibs below 4 mill

Firstly the Fibonacci definition for Learning J throws a curved ball 
near the end:

fib=: u ^: (v1 `v2)

Generally I understand that the gerund lets me pass multiple arguments 
to a function which only
takes a single, but what is it doing here?

Well, I did the crime of continuing without understanding that, so 
here's what I arrived at

+/ (#~ <&(4*1e6)) (((0=2&|) # ]) fib 40x)

And I think that's almost acceptable. My two main complaints is that I 
have 2 parenthesised functions
lined up as a sequence-parser and I imagine that they would be coupled 
some way, ie

+/ (take while < 4 mill and only the evens) fib 40x

Secondly, I couldn't figure out how to just let an infinite amount of 
Fibonacci numbers return, so I hardcoded
40. fibs^:a: gave me a domain error for some reason.

#3: Largest primefactor of 600851475143

I'll admit, I felt like cheating on this one

 >./q: 600851475143x

And thats as far as I got, because this simply makes no sense to me :)

((-:|.)@:":)"0  NB. Tests for palindrome


Thanks,
Lau