I'd noticed that from SAJ.ijs. Actually my example is a bozo one, just for illustration. Since I've had such bad experiences of this nature, I take special care when coding logical tests.
But I've seen some strange side-effects from writing =: for =., especially with "y" involved, that I resolve to go on proactive bug-hunts in my code, looking for all instances of y =: and =. Maybe I'll adapt LoBrow to show big red warnings. (There... it'll be of some use after all!) Ian On Fri, Jan 22, 2010 at 1:27 PM, Dan Bron <[email protected]> wrote: ... > PS: Unfortunately, for your use, the SAJ script takes specific pains to > allow the programmer to shadow a global name with a local name (as usual), > even if that global name is SAJed (constant). So if you tried > > if. replyYES=.y do. ... NB. Note =. not =: > > J would happily execute it... ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
