scaled_iota name

[jtlap]

iota is a good name as it is the same as in std, but scaled_iota is not.
When I read scaled_iota I think of iota(...)*scale and that's not it.

proposals
arithmetic_progression : a bit long
.ap : a bit short
arith_prog or arithmetic_prog are midway

progression can also be replaced by sequence or seq.
It is also possible to write new functions on the same model geom_seq harmonic_seq...

I thingk I prefer arithmetic_seq
What do you think ?

[DenisYaroshevskiy]

It is not iota(...)*scale ??

I thought it is.

[DenisYaroshevskiy]

Well - the base can be adjusted - but that's also the case with the regular iota.

[jtlap]

the code is ok it's the name that is not scaled _iota is what you mean only if the first term is 0.
it computes base+scale*(n-1) which is general arithmetic sequence and scale is not a "scale" but a "step"

[jtlap]

in term of algorithm it's exclusive_scan vs inclusive_scan in a peculiar case

[DenisYaroshevskiy]

the C++ standard iota and standard iota view allow for a different initial value

https://en.cppreference.com/w/cpp/ranges/iota_view
https://en.cppreference.com/w/cpp/algorithm/iota

I think it should be called smth iota because the user should be able to find it.
If we call it not iota, user will have troble.

[jfalcou]

I would prefer iota_step as I remember being confused by the scaled suffix when looking at the code and not seeing any scaling happening.

[DenisYaroshevskiy]

Ok - I see it confused ppl. Hmm.

The word iota should be and should come first. iota_step - eugh.

iota_with_step? Maybe?

[jfalcou]

iota_with_step soudns OK to me

[jtlap]

for me too.
Who does the job ?
I can if you want

[jtlap]

perhaps tomorrow...