@Adám What?! So under ⍢| is storing phase information? ⍨

@B.Wilson ⍢ doesn't exist yet in Dyalog.

@Adám Yeah, I meant in whatever is executing the example phrase there, be it your brain wetware or some other hardware.

@B.Wilson Ah yes, but that's pretty trivial in the head.

Is storing auxiliary extra-array information within the execution context of Under something actually being considered?

@B.Wilson Implementation details have not been discussed.

I think, arguably, the spec for orf in Phase 2, Task 5, should return vector of vectors of strings instead, since FASTA format can potentially contain more than a single sequence.

> takes a right argument that is the name of a file containing a single DNA string in FASTA format

@user21102602 Hi Jimmy Brown, if you want to participate here, please email access@apl.chat

Just curious, but what's the reason for the name class jump from 4 to 8?

@B.Wilson I don't actually know, but my assumption is that ¯1…4 are the traditional APL ones, and Dyalog's additions then descend from 9.

I'm proposing 11 12 13 14 19 for system constants, variables, fns, ops, objs.

:63189792 Wait... Ninja delete.

Is the 5..8 stuff .Net?

No, my mistake, I somehow mixed up sub-classes.

@B.Wilson I've written an email to Geoff Streeter asking for the reason.

@Adám Wait, so the "system" adjective there applies to the entire list? What kinds of thing is a system constant?

E.g. ⎕A

Ah, okay.

Also ⎕TS even though it changes a thousand times a second. But you cannot assign to it (I always wondered why), so it is considered a constant.

⎕AV can be changed by the user, but only as a side effect of changing the system variable ⎕AVU, so ⎕AV is a "constant" too.

Heh. TIME IS CONSTANT.

So is your name, ⎕AN

Fun fact: There are no monadic system operators in Dyalog.

Huh. That's mildly intriguing.

Pretty cool that Geoff Streeter is still working on Dyalog.

I think there should have been one: ⎕TSPAWN rather than &

@Adám Why would you prefer such a verbose token?

Because it isn't an array manipulation/computation thing at all. It is very much a systemy thingy. Also, its inverse is called ⎕TSYNC, not e.g. $

That's reasonable. Though by that argument, → also seems half way in that camp as well.

Indeed. Today, we have :GoTo and :Return (because → really should be considered syntax, not a function proper.)

I've lately also begun to prefer ⎕OPT over ⍠ since, again, it is a systemy thingy. In fact, it currently only pertains to system functions (and .NET type casting).

@B.Wilson Yeah, until April 6.

So you like having clear visual boundaries between these "roles" that the tokens play?

Yeah, or at least don't hog precious glyphs for non-core things.

@Adám Wait. He's retiring?!

He is. The first person to retire from Dyalog ever.

(I'm not sure if that's good or bad.)

Wow. That's both surprising and pretty momentous.

Everyone else has just moved on (from this life, or to a different job).

People get poached not very infrequently?

Not often, but it does happen. I was also offered an APL job for about 1.5×my current salary.

@MasterQuiz OK, we can consistently repro your issue now, and have logged it as number 20416: Crash in Unicode IME after using Clipboard history to paste in text, get an error and then clicking Desktop and back in Dyalog. Thank you so much for reporting it.

@Adám Are glyphs precious mostly because minting new ones is so hard?

:-) There's limited keyboard space — and brain capacity.

Tell that to China or Japan! :P

Huh. I just realized that result shyness can depend on arbitrary computation within the function.

Yes, in a dfn: Shy←{⍺:_←⍵ ⋄ ⍵}

Yeah, exactly. {?2:_←42 ⋄ 42}

Okay. So the ⎕PATH mechanism only works for functions and operators. Trying to ⎕EXPORT a variable is just a no-op.

@B.Wilson Don't use ⎕PATH:

> ⎕PATH may therefore be considered an archaic feature.

You beat me to it. Was going to ask why the feature is effectively deprecated.

It cannot be fixed sensibly while remaining backwards compatible.

Gotcha.

For 2:2:3, if I have an one-liner but takes 2GB ws to pass the example test and a more "practical" one that makes lot of use for → and procedure style programming, should I consider submit both or the functional one would be preferred, or the answer would only be judged as valid if can pass the example tests in default setting?

@LdBeth We stress "the appropriate use of array-oriented techniques"... a one-liner solution that consumes vast resources (space or time) probably won't fair as well in judging as one that is procedural but runs using a fraction of the resources.

Got it.