Sounds like an old wives tale to me. On Mon, Aug 8, 2011 at 5:12 AM, Phil Smith III <li...@akphs.com> wrote:
> I sure don’t (from a comment to > http://queue.acm.org/detail.cfm?id=2010365):**** > > “In the '70s, IBM released a toolkit library for assembler programming on > the 3270 mainframe, under VM. One of the routines was an assembler snippet > which you were supposed to put at the end of programs to signify a clean > exit: it was something like, "set register 17 to 0; quit" -- two assembler > instructions. So programmers throughout the '70s were adding the macro for > this to the ends of their programs. Turns out, IBM put a bug in it, so that > it set the wrong register (a two-instruction program with a bug in it) thus > virtually all assembler programs on 3270s written within a 10-year period > have this bug in them: IBM was finally forced to modify VM to cope.”**** > > ** ** > > Of course the “register 17” is suspicious (as is “the 3270 mainframe”)…is > this a confused reference to IEFBR14?**** > > ** ** > > …phsiii**** >