I believe it is still the case that some text calculations are more efficiently managed by working with integers that represent literals. I have not yet had such a situation but I'm sure I've seen some code from Roger Hui that did this. It may have been to manage Unicode text without the complications of varying character byte lengths.
Taking irregular data and finding a way to make it regular is one of the first instincts of an array programmer. ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
