Mastodawn

So, in Lispy languages you can access the head and tail of a list with `car` and `cdr` (insert historical sidebar here about what these names originally meant). But maybe you want the second element of a list -- that is, the head of the tail of the list. Or maybe the list's elements are themselves lists and you want the tail of the head. And so on. For this reason, there are pair accessor shorthands: (caar ls) is shorthand for (car (car ls)), (cdar ls) is shorthand for (cdr (car ls)), and so on.

Lindsey Kuper Mar 4

In the Racket docs (https://docs.racket-lang.org/reference/pairs.html#(part._.Pair_.Accessor_.Shorthands)), these pair accessor shorthands go all the way up to four levels of nesting: caar, cadr, cdar, cddr, caaar, caadr, cadar, caddr, cdaar, cdadr, cddar, cdddr, caaaar, caaadr, caadar, caaddr, cadaar, cadadr, caddar, cadddr, cdaaar, cdaadr, cdadar, cdaddr, cddaar, cddadr, cdddar, cddddr.

BUT WHY STOP THERE?!? Why can't I write `caaaaaaaaar` and have it just work?

4.10 Pairs and Lists

Lindsey Kuper Mar 4

Therefore I propose that whenever the programmer writes c(a|d)*r, the language implementation ought to generate the appropriate nested sequence of car and cdr calls, no matter how deeply nested. I'll buy a drink for whoever implements this

Lindsey Kuper Mar 4

oh, I guess it should be `c(a|d)+r` lol

@lindsey `cr` as notation for the identity sounds fine actually

∴ esoterik ∴Mar 4

@lindsey nah, i like `cr` as a noop

Marcos Dione Mar 5

@lindsey now you got two problems :)

James Widman Mar 4

@lindsey when we say these things out loud, listeners should use rhythm to determine the number of `a`s, with tempo to be determined by syllables-per-second leading up to the utterance of an accessor

Lindsey Kuper Mar 4

@JamesWidman https://social.belkadan.com/@jrose/statuses/01KJX069EHK0WEBS8JY9TX0H8R

Jordan (@[email protected])

@lindsey You would appreciate this, from the professor I TA’d for: https://people.eecs.berkeley.edu/~bh/61a-pages/Lib/pronounce.scm

belkadan.com

Daphne Preston-Kendal Mar 4

@lindsey https://docs.racket-lang.org/cadnr/index.html

c(a|d)ⁿr

Lindsey Kuper Mar 4

@dpk YES

the author is entitled to 1 (one) drink

Lindsey Kuper Mar 4

@dpk I assume this behavior is going into the R7RS standard

Tim Ward ⭐🇪🇺🔶 #FBPE Mar 4

@lindsey That's called Claude, isn't it?

Alex Rudnick Mar 4

@TimWardCam @lindsey I have bad news for you about LLMs and correctly counting the number of times a letter occurs.

Paul Lalonde Mar 4

@TimWardCam claudr

Tim Ward ⭐🇪🇺🔶 #FBPE Mar 4

@Flux 🤣

Peter Ludemann Mar 4

@lindsey Haskell already has this (derived from Miranda), as does Prolog.
Also: https://www.cs.kent.ac.uk/people/staff/dat/miranda/wadler87.pdf

paulette d. koronkevich Mar 4

@lindsey @wilbowma @samth

Lindsey Kuper Mar 4

@koronkebitch @wilbowma @samth apparently it's been done (of course): https://chaos.social/@dpk/116172269420528793

Daphne Preston-Kendal (@[email protected])

@[email protected] https://docs.racket-lang.org/cadnr/index.html

chaos.social

@lindsey You would appreciate this, from the professor I TA’d for: https://people.eecs.berkeley.edu/~bh/61a-pages/Lib/pronounce.scm

Abhijit Menon-Sen Mar 4

@lindsey Some say the implementers are still trying to count the number of "a"s in "caaaaaaaaar" and getting different answers each time.

@lindsey I did the moral equivalent of this in Smalltalk, and was told by experienced Smalltalkers that this was not really the intended use of doesNotUnderstand

Martin Escardo Mar 4

@lindsey This is so clearly readable.

David Beazley Mar 4

@MartinEscardo @lindsey Can this be mixed with regex somehow?

Lindsey Kuper Mar 4

@dabeaz @MartinEscardo I personally find functions like `cdadr` to be charming and delightful. At the same time, I would never, ever want to subject anyone else to them. They are truly blursed.

Martin Escardo Mar 4

@lindsey writes "At the same time, I would never, ever want to subject anyone else to them."

Sincere thanks for that!

Paul Hoffman Mar 4

@lindsey @dabeaz @MartinEscardo I wish someone had used the word "blursed" when I was forced to learn those functions exactly 50 years ago. It would have made me not feel as stupid for being so confused about them. (This was for OG Scheme.)

David Beazley Mar 4

@lindsey @MartinEscardo I remember taking a SICP group out for a coffee break (when I still taught in person) and we were all standing in line talking about some car-this, cdr-that, cadr thing. We got a lot of weird looks for that.

Martin Escardo Mar 4

@dabeaz writes " We got a lot of weird looks for that."

This reminds of a story from the last millennium.

I was in the queue with various people for lunch, at Kings Buildings in Ediburgh, including Alex Simpson, John Longley, Samson Abramsky and Gordon Plotkin, and I was saying verbally something about "continuous functions from X to two".

Then somebody asked "what is 'two'". And I said "one plus one", which was clear to them (they wanted to know whether I was talking about the Sierpinski space or the two-point discrete space).

But we got, repeating your words, "weird looks for that" from everybody else in the queue.

Omar Antolín Mar 4

@MartinEscardo @dabeaz @lindsey That story in turn reminded me of the time I was explaining Conway's field structure on the natural numbers to a friend, but I only knew a very basic definition. I knew the addition was "nim-addition": write the numbers in base 2 and add the corresponding digits mod 2 without carries. And I knew that the product of two natural numbers is the smallest it could be, that is the smallest number for which you can't get a contradiction by assuming that's the product and using the field axioms. This is a notoriously impractical definition of multiplication. We were walking and talking, trying to build out the multiplication table one product at a time. It tools us a long time to figure out 2×4. We proved it couldn't be 0, it couldn't be 1, it couldn't be 2, etc. Just as we were arriving at a taco stand, we figured out that in this field structure 2×4=8. We got funny looks from the taqueria and the people eating tacos for being so damn excited about 2×4=8.

Omar Antolín Mar 4

@MartinEscardo @dabeaz @lindsey Jesus Christ, that must have been something like 25 years ago. 😱

@lindsey I once defined a `cr` macro so you could write `(cr addad x)`.

Leah Neukirchen Mar 4

@lindsey my private lisp uses cr for the identity function :)

Lindsey Kuper Mar 4

@lindsey People usually need to iterate over a list, but some direct access is not a bad idea, sometimes. I hear about these for a first time.