dhblah

I mean, there are several tricks along the way.

@Ixrec hi, remember I've been asking about binary search. I finally got the idea. I.e. converted 3 predicates solution to a single predicate solution. Also I have a sort of step by step conversion instruction. Planning to make a page on github on it.

and I just was a bit distracted and lost a thread

I was just looking at binary search again today

hi

got to go to work now

Thanks a lot, you've been really helpful

well, I still don't get it, but thanks for the clue, I'll look at it

oh, you mean lo = mid +1/hi = mid - 1

what do you mean by +1/-1 symmetry?

it works, cool

let's add + 1

I mean, Imagine I was the guy who devised an algorithm, it's like, wow, my algorithm doesn't work

ye, I just don't get it

Thanks a lot for your help Ixrec, I'll check it when I have free time. thx bye

The code will get stuck in a loop. It will always select the first element as mid, but then will not move the lower bound because it wants to keep the no in its search space. The solution is to change mid = lo + (hi-lo)/2 to mid = lo + (hi-lo+1)/2, i.e. so that it rounds up instead of down.

i.e. this one

also that + 1 bit breaks my mind

I guess I just need to take it and slowly transform to a second one

the only solution I can clearly reason about, is the first one

it kinda works, but I don't understand why

but I don't understand it

look, I mean I can write that binary search even in predicate version

and pass just predicate and domain

or even probably combine function with predicate

you can also pass a predicate instead of target

and a function domain

so in this algorithm you described, you basically pass in it a function

Ixrec, yes, that's the algorithm I had in mind.

yes, I'm talking about taking first algorithm and replacing element addressing with a function call

That's just it

Ixrec , I'm talking about replacing element addressing with a function call, because one of the arguments I heard for the second solution was that you can use it not with an array, but just with a function. So my counterpoint is that you can also use first solution with a function as element addressing is basically a function too.

I have two points in defence of the first soltuion: 1) you don't need to provide a complete search space (i.e. filled A array), as you can easily replace element addressing with a function call. 2) it's easier to reason about.

what it actually does inside it

the problem with that solution is that it's hard to reason about

I"m saying that you can use a function instead of addressing element in an array

you can replace A[mid] with function

topcoder.com/community/data-science/data-science-tutorials/…

the second algorithm is binary_search(lo, hi, p)

and it doesn't help that second algorithm is pretty hard to reason about

that's why I'm confused

because I'm trying to understand all the quirky bits in the second algorithm

I'm confused

lol

you devised

you can interchange operator for a method call

@Brandin 1) algorithm is not the same. 2) you can replace A[mid] with a function call, so you don't actually need an array as a parameter 3) you can replace '==' and '<' operators with your own type 'methods' 4) as for the boundaries, probably there is a way to implement this functionality with the first approach too.

se

doesn't make sene

but the second one