On Wed, 22 Feb 2017 15:05:54 -0600, Paul Gilmartin ([email protected]) wrote about "Re: Question about PARMDD" (in <[email protected]>):
[snip] > This thread began with a discussion of the PARM string. I'd like to think of > JCL > as source code, not as data. You can consider JCL as whatever you wish. The point I was making was that the PARM string is data, not source code. > And I definitely think of JCL as sorely deficient > in string manipulation operations. It was developed a half-century ago, > subject to severe resource constraints. I first coded OS JCL under MVT more than 40 years ago. I coded DOS JCL before that, and OS JCL is luxury in comparison. I also coded GCOS JCL on Honeywell mainframes in the early 1980s; OS JCL is *serious luxury* in comparison to that, That said, I will not defend JCL's limitations. Its string handling is largely limited to symbolic substitution using syntax taken from the macro language of System/360 assembler. From the point of view of the average COBOL programmer it is almost impenetrable. Ideally, JCL should have civilized syntax for string manipulation and for flow control. This would include SELECT/WHEN/OTHERWISE, DO-loops. WHILE-loops as well as IF/ELSEIF/ELSE. The string manipulation would include conditional substitution (ternary operator?) as well as concatenation, extraction and removal of substrings, scanning strings(*) for patterns and a whole lot more. We're basically talking Python here, and a bit more besides. (*) I think regular expressions would send most mainframe programmers scatty, but it would be nice to have. -- Regards, Dave [RLU #314465] *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-* [email protected] (David W Noon) *-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-*-* ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
