On Fri, 18 Jun 2021 14:02:19 -0400, Bob Bridges wrote: >Almost forgot this one. My abandonment of one-byte variable names isn't in >pursuit of better self-documentation (though it does contribute a little to >that); it's because once in a REXX program I had a variable X and tried to >concatenate it to a string: > > result='string'x > I think the trailing string qualifier is counterintuitive. MFC may have chosen it to simplify lexical analysis.
In early IBM Rexxen 'B' was not a striong qualifier. Nowadays it denotes "Binary". The change seems to have broken some code, for which there's an apologia in the TSO/E Rexx Ref. Likewise the introduction of 'R' as a delimited string qualifier in ISPF Edit must have broken some code. C's use of leading '0' to indicate octal values is an abomination. >And for all loop variables I use not i-something -- I tried that at first, >but there are too many two-letter reserved words beginning with 'i' (IF, IN, >IS) -- but j-something: jr for a row, ja for an ACID, jp for a profile, >whatever. Again, it's a small help, but it works for me. > Eek! Hungarian Notation! -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
