`list.windowed(w).map { it.sum() }.max()`

# random

[karelpeeters]

list.windowed(w).map { it.sum() }.max()

[voddan]

.maxBy {it.sum()} ☝️

[karelpeeters]

That doesn't return the actual sum though, only the original list.

[voddan]

Right

[karelpeeters]

I've been bitten by this a couple of times before too 🙂.

[karelpeeters]

.maxBy {it.sum()}.sum() 🧌

[voddan]

their function name MaximumWindow is misleading 😠

[karelpeeters]

"CS people"

[voddan]

And now I am stuck thinking which one is actually using the least memory...

[karelpeeters]

Yours, mine should have been lazy.

[karelpeeters]

list.toSequence().windowed(w).map{ it.sum() }.max() would use O(1) memory for the max part I think.

[voddan]

It allocates at least w bytes for windows itself , so it is sub-optimal already

[karelpeeters]

True, but here weren't any requirements on the memory consumption of the windowing step.

[voddan]

the O(1) is a requirement on memory. The optimal hand-written algorithm doesn't need arrays at all