On Wed, 2 Sep 2020 18:42:19 +0000, Seymour J Metz wrote: >> However, if IBM were to extend in that fashion it would be at the >> cost of rendering possibly existing user code "bad form".
>No; there's a reason that I wrote "knowingly". You're describing a situation >where the user *doesn't* know that IBM will be adding that keyword. > So if a naive user unwittingly uses a keyword as an identifier regardless of chronology (there many possibilities) it's not bad form although a more sophisticated user may perceive it so? Or do you insist that a competent programmer will have memorized PL/I's entire vocabulary? It's very subjective. Should the compiler in pedantic mode issue informative messages on such errors in form even as GNU C compiler flags "if (A=B) ..."? Back to FORTRAN and blanks, a colleague was aghast to learn that blanks matter to the nascent PL/I. But she probably didn't understand that two blanks would be no different from one. ________________________________________ On Wed, 2 Sep 2020 15:02:51 +0000, Seymour J Metz wrote: >PL/I has no reserved words, so IBM can extend the language without breaking >existing code. However, it is bad form to knowingly use a keyword as a >procedure, label or variable name. -- gil ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
