I have used REXX and J for many years, in a variety of environments. I agree with Boyko Bantchev that REXX is a remarkably effective scripting language. I also agree that its data manipulation tools are impressively useful, almost despite the minimalist nature of the language. But, I don't believe REXX can touch J's capabilities with problems of any complexity.
I just wrote a REXX script today in z/OS Unix to monitor a SSH server and email me status reports hourly. I'm sure there are a multitude of ways this could have been done, but I found the REXX solution quite simple and even somewhat elegant, thanks to the ease of the 'address' mode of accessing many diverse APIs, singly and in groups. Today I also used J to analyze a 50 MB log file from the same SSH server. I have used J for scripting in other environments. I have also used J for complete, medium scale applications, which I would never have tried to write entirely in REXX. One element I find essential with both languages is that they both have a large base of freely available routines. I depend on these libraries and collections to enhance both my productivity and my understanding. The single feature I have enjoyed (and cursed) most with APL and J is that, in my experience, APL and J are indeed tools of thought. With most languages, I find that programs almost write themselves. With J, I am often not truly productive until I spend the time to immerse myself in the problem, the data and the language, when critical insights will then bubble up and drive my thinking in totally unexpected directions. -- cheers, David Mitchell ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
