Fraser Jackson wrote:
> I wanted a general purpose adverb so omitted them on the grounds
> that they were designed for nouns of the same type.
Of course, general purpose adverbs have been proposed.
For example, this one was posted earlier:
seq=: 1 :(':';'>u&.>/(<"0|.y),<x')
0&(+seq)\ i. 9
0 1 3 6 10 15 21 28 36
This is general in the sense you've specified.
I do think there's a place for induction from a base case with values
in an array supplying data for each inductive step, and where those
values are not in the domain of the base case. That said, J currently
requires that you map those values into the base case's domain. The
above definition is an example of this kind of mapping.
But I wouldn't call the above "fully general" because there are
potentially an infinite set of such mappings, and the above case
only deals with one of them. These issues can become relevant
when you near the resource limits of the machine.
Then again, I don't think that any real computer program can ever be
"fully general". And, I think you get the best system when you pick
the specific generalities which are relevant to the issues you are
currently dealing with.
Put differently, good programming tends to eliminate irrelevant
generalities from the system.
--
Raul
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
