>simplicity. Simplicity is a virtue?
On designing econometric models one of the best advised: "Keep it
Sophisticatedly Simple." I think his advice also applies to programming;
actually that is the primary reason why I prefer J versus other
conventional programming languages and even to APL (at least the decade old
versions of APL that I used to know). Moreover, writing only tacitly keeps
the list of concepts simple in my mind.
To stress the main point of my initial message; the case I am making for
recursion without $: is, for instance, the fixed tacit (or otherwise)
implementation of the adverb sna. The (atomic representation of the)
Recursion adverb is an interesting testing argument taken by sna.
Nevertheless,
I posed the implementation of Recursion as a puzzle and yours and any other
variants are very welcome.
Recursion=:1 :0
'`if else test next'=. m
[: (if else^:(-.@test))/f. (, next@{:)^:(test@{:)^:_ f.
)
fact=. *`1:`*`<: Recursion
fact 5
120
So far so good. Unfortunately,
fact 0
0
Also,
Fib=. >&1`(i.@>:)`(] , +/@(_2&{.)@])`<: Recursion
Fib 7
|domain error: Fib
| Fib 7
|[-15]
Can you fix your interesting (primitive recursive) version?
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
