[seqfan] Re: most obscure OEIS code ever?

[Joerg Arndt]

* Douglas McNeil <mcneil at hku.hk> [Jul 01. 2010 14:55]:
> %o A178683 (J programming language, see http://www.jsoftware.com)
> %o A178683 . aXXXX =: (5 #. ] {::~ [: (i. >./) */&>)@:part"0
> %o A178683 . part =: 3 : 'final (, new)^:y <<i.1 0' NB. Here & below due to Hui
> %o A178683 . final=: ; @: (<@-.&0"1&.>) @ > @ {:
> %o A178683 . new =: (-i.)@# <@:(cat&.>) ]
> %o A178683 . cat =: [ ;@:(,.&.>) -@(<.#) {. ]
> 

sendmail.conf on steroids, I am thoroughly impressed.


> There are a handful of other sequences with J codes, but this is
> probably the most impressive.  Surprisingly, J appears not to be a
> parody esolang, which was my first guess-- or if it is, the authors
> have admirably moved all-in on the joke and have kept their tongues in
> their cheeks for almost twenty years.  One's tempted to suspect a
> drunken bet gone horribly, horribly wrong..
> 

Well, it's a concise language indeed,
the following (supposedly) computes A005249:

 H=: % @: >: @: (+/~) @: i.
 det=: -/ .* 

Not sure about readability  ;-)

http://en.wikipedia.org/wiki/J_(programming_language)
cf. also http://en.wikipedia.org/wiki/Tacit_programming
" ... or more pithily as pointless programming."

Maybe this really is an in-joke of epic proportions.


>  [...]