Closed dylannrees closed 6 years ago
The first four should probably look a bit more related. Since we recently added Œœ
for odd-even, I'd say we use Œo
and Œe
for odd and even, and consequently Ðo
and Ðe
for the corresponding quicks.
The last two sound useful, but I wonder if a single-byte not quick would be worth it. It would vectorize at the same depth as the underlying link, so <ɲ
and >ɲ
would be ≥
and ≤
, which we also do not have as built-ins.
Caught a bug in even indices quick and made the changes you suggested.
I don't think the not quick ɲ
vectorizes at the depth of the link it consumes since I used the ¬
atom to do the call. How can that vectorization behavior be accomplished in a quick?
I'm not sure about the ɲ
quick yet. Could you keep the first four for now, so I can merge them?
Ðo
Apply to odd indices.Ðe
Apply to even indices.Œo
Get elements at odd indices.Œe
Get elements at even indices.Œḃ
Bit complement.(z+1) % 2
ŒỊ
Not insignificant.abs(z) > 1
The last two would only save a byte if they need to be a single link (
ḂC$
,ỊC$
), but I've found myself use each of those a couple times.