RE Boss asked:
> why "reversing y along each axis" was
> introduced in defining ;.0
My quick search didn't turn up a rationale, but in response to your (RE Boss')
"Reversing all ranks" thread, which I linked to earlier, Roger had this to say:
Ken told me that sometimes, when he came up with a neat
solution to a problem, he'd be accused of putting in
a special construct in the language just for that problem.
(And that was why the solution was so neat.) But the
accusation lost its sting when it was repeated in problem
after problem, in multiple diverse fields.
(From
http://www.jsoftware.com/pipermail/programming/2005-December/000322.html )
As I read it, it was posted more-or-less in defense of the definition of ;.0
, so maybe Ken saw this (reversing all ranks) as a recurring problem. Roger,
can you shed any light?
Actually, RE Boss, maybe you can. You started that thread by asking for a
replacement for |."3@:(|."2)@:(|."1) . Do you remember why you wanted to do
that? I personally have not seen this problem recur in field after field.
In fact, after some research earlier today, it seems the recurring "all ranks"
problem is in fact the one that started this thread: padding an array on all
dimensions. Trimming (the opposite of padding) an array on all dimensions came
in second.
-Dan
PS: I don't remember ever wanting to "reverse all ranks", but I have wanted an
"alternating reversal" sometimes. That is, I wanted to reverse every other
item of an array. One of my prouder J moments was discovering (A."_1~ _2 |
i.@:#) for this purpose. There are other ways, too:
http://www.jsoftware.com/pipermail/programming/2006-November/004198.html
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm
