the order of arguments

stevan apter — 2000-06-19 22:48:56

fellow concatenators:



in joy:



    joy>5 3 -

    2



but as i've implemented the primitives in conk:



    k>5 3 -

    -2



i prefer the conk interpretation, since [3 -] is then the program which

subtracts the item on the top of the stack from 3, which seems more natural

than the program which subtracts 3 from the top of the stack.



this holds for division, >, <, >=, and <=.  i can't think of any others

offhand.



this wasn't intentional, except insofar as i want the conk primitives

to be mapped by one rule to the k primitives.  if i reverse the arguments

of -, then i would have to do that everywhere, and the resulting set of

primitives is extremely cumbersome.



unfortunately, it means that several of manfred's joy programs won't run

without a bit of surgery.  (i discovered this the hard way, thinking that

my implementation of condlinrec must be wrong, since none of the examples

would run!)



any opinions?



--



ps:  the latest version can be found at www.kx.com/technical/contribs/stevan.