The following code works, and solves the following problem: identify
the items in a boolean table that contain non-contiguous ones.
negateHits=: * (-. + -)
ix10idot =: 1 i."1~ 1 0 E."1 ]
ix01icolon=: (] negateHits (] = {:@[EMAIL PROTECTED])) (1 i:"1~ 0 1 E."1 ])
b=: #: i. 2^3
(ix01icolon > ix10idot) b
0 0 0 0 0 1 0 0
I'm very happy with negateHits, although it does have a data-control
x-y parameter arrangement rather than the preferred control-data
arrangement.
The two index-making definitions are closer than they look. The
index-of-first-occurrences doesn't need to have the extra code because of the
directionality of greater-than in the final line.
It seems to me that the next pass here, if this were to be refined,
might be to create a conjunction that would take the bit pattern (1 0 or 0
1) and the associated verb (i. or i:) then produce a verb. I don't
think it is worth doing for this problem, but perhaps my skill is limiting
my vision.
Also, if I missed a technique of solving this problem more simply than
I have, I'd like to know. Thank you, in advance, for any comment!
Tracy
____________________________________________________________________________________
Be a better friend, newshound, and
know-it-all with Yahoo! Mobile. Try it now.
http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
