Given how long APL has been around, it occurs to me that lots of hand-written APL probably went on in the early days. Now I'm curious whether there were cursive forms of the squiggles.

are tacit definitions allowed in the RIDE editor at all?

@B.Wilson Not that I know of. Hand-writing APL is still common, especially at various meetings and workshops. Aaron and I were just hand-writing APL last week, and he just uses normal forms, even when otherwise writing in cursive.

@south Yes (although RIDE isn't really an editor).

Am I misremembering something, or was there a proposal(/dialect implementation) for ⍴ to take -1 as one axis and have it compute the necessary size?

I thought Vision did that, but apparently negative is reserved for reversed axes while it uses ugly 0.5/1.5 for autosize (and neither do what I want, which is an error if it can't be filled exactly)

oh, Extended does what I thought

@Adám, why the switch?

@RubenVerg That's correct.

@RubenVerg To gain the functionality of reversing. I could add 3.5 to error.

idk, something about using .5's looks ugly to me

but also, as much as reversing axes might be nice, it does go against reshape being just take on the cycle of the ravel

(that is, ,s⍴a is (×⌿s)↑(,a),(,a),(,a),...)

@RubenVerg what were you expecting to happen if can't fill - recycle values?

@Silas I would expect, say, ¯1 2⍴⍳3 to give me a domain error

@south If you mean use )ed to edit tacit function, you can do that by define a nildic tradfn that returns a tacit function as result: github.com/LdBeth/pnmaster/blob/main/APL_code/Image/shift.aplf and that definition can be used usual tactic definition (although if the definition contains side effect it would be evaluated every time so there is a slight difference)