BASIC was not supported on a SYSTEM/3 machine I don't believe, but there was a BASIC reference manual 65-2211-7 1974, that would have been for the 370, same time period at least.
On Wed, May 7, 2025 at 8:13 AM Steve Lewis via cctalk <[email protected]> wrote: > > I wonder if you might be making unwarranted assumptions about the extent > of what is encompassed by the emulation. It’s at least plausible that the > > state of the emulation required for implementing the APL interpreter was > missing some non-trivial set of features required by the BASIC interpreter. > > Certainly possible - we've found the jump tables in both emulators that > gives an inventory of what opcodes they supported, and to fit in the > capacity and cost (of what amount of software could be fitted into a > slotted card at the time) that they were targeting, some compromises had to > be made [ so neither were a "general 100% cycle accurate" kind of > emulation, rather the focus was on supporting the essentials necessary for > those respective applications of APL and BASIC, adapted from line-printer > over to kbd/CRT fashion ] > > I recall in the SCAMP Joe George engineering docs, there was an entry where > they had trouble funding or obtaining a "spare" System/3 for their > development (that being IBMs own internal developers). I'm on travel and > will have to lookup the exact reference later. Perhaps similarly, whatever > S/360 they borrowed or used previously was no longer available. So > (maybe) just being pragmatic on what was available to them at the time, > they went the S/3 route. > > Another reason might possibly be: iirc, 1975 the trial portion of IBMs long > antitrust suite was starting (the decade long one ending right at the > release of the IBM PC in 1981). That being the issue of coupling software > to their hardware. A thought was that by at least showing the concept that > "some other language" could essentially be plugged into the system, it at > least gives the impression that a third-party could technically do the > same. Of course no one ever did (I'm not sure if they'd have access to > enough info then to do so), but there was mention of a FORTRAN card concept > for the system. Aspects along these lines are also implied in the dev > journal mentioned above (if not for third party, but even in internal > debate and justification by Friedl to get the project funded, a la "you > want to build WHAT?" kind of discussions to a mainframe oriented business > unit). Dennis Robinson might still be available to chime in on this > aspect. > > -Steve > > > > > > > > > > > > On Wed, May 7, 2025 at 5:18 AM r.stricklin via cctalk < > [email protected]> > wrote: > > > > > > > > On May 6, 2025, at 11:32 PM, Steve Lewis via cctalk < > > [email protected]> wrote: > > > > > >>>> OS/360 TSO BASIC, pgm-nr 5734-RC2 round 1971 > > > > > > Interesting - then I wonder why in 1974 they bothered to build a S/3 > > > emulator into the IBM 5100. They had already done the S/360 emulator > > and > > > gotten APL working, so why not use this OS/360 version of BASIC? > > > > I wonder if you might be making unwarranted assumptions about the extent > > of what is encompassed by the emulation. It’s at least plausible that the > > state of the emulation required for implementing the APL interpreter was > > missing some non-trivial set of features required by the BASIC > interpreter. > > > > It’s also worth considering, as you point out, that perhaps the BASIC > > inerpreter depended on some OS services. APL\360 had originally been its > > own standalone timesharing environment. I’m unsure of whether that > > situation was fully at an end by the time of APL\360 on DOS and MVT > > (1969ish) but I don’t doubt the standalone code was likely still > reasonably > > fresh by the time work was underway on the 5100. The S/3 BASIC may have > > been in a more reasonable state for adapting to standalone operation on > the > > 5100. I haven’t kept up with all the scholarship around the history of > the > > 5100 but it’s also not exactly implausible that there was some phase of > the > > project where the thing had been two separate products. > > > > ok > > bear. > > > > >
