I sent off to Quote-Quad a paper that uses J (and APL) notation for showing
proofs (no guarantee they'll accept it after the Iverson volumes).
Some of the things I do are easier in APL than J because APL won't let me
do some of the things J permits. On the other hand, in J, if I pull things
all together I run afoul of some of the constructs, in particular if I
have a matrix X and I want the ith column of X against all X in
+/ .* I get messed up if the i is part of an earlier +/ .* so that i, too,
runs across all columns.
I suppose if I break it into several verbs it might be simpler, but
relatively inscrutable, while totally insrutable if I don't do some
indexing.
Finally I am using >./ .- and the like in the inner products, not +/ .*
Ralph Selfridge
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
