Re: Fwd: Speed versus clarity

On 2023-03-29 12:04, Bill Allombert wrote:
> For example for issquare(N), in some application, N wil almost always 
> be a square,
> while in other N will almost never be a square. Ideally one should not use the
> same algorithm for both case.

Then maybe issquare wants to grow a hint-flag (as an extra parameter),
to give the caller a way to influence that decision.

Another option is to have 2 names:
issquare_while_unlikely
issquare_while_likely

but then what to do if
issquare_while_unlikely(t_COMPLEX)
gets called ...
:)

-- Ruud