SCAMP: A Review 50 Years Later — voidstar
Had a fascinating opportunity to explore SCAMP, the prototype to the IBM 5100 !
4 Likes
Very nice. Thanks for the pointer to the extra detail in your VCF thread.
Hope you don’t mind me pasting one of your images, the minimal front panel:
We should perhaps also link to your earlier thread with your presentation:
IBM 5100 discussion at VCF SW 2023
And an earlier thread on your explorations:
All about the IBM 5100 from 1975 (also IBM 5106 and IBM 5110)
3 Likes
Thanks, perfect consolidation there!
I had hoped to speak more about it at VCF in September, but the info came a little late for me to get prepared and booked in time. But still wanted to get the info out there during this anniversary year.
I’m off and on between travel, but do plan to have another batch of updates within the next week as I organize the remainder of pages.
1 Like
Very nice article – thank you!
Update #1: I’m about halfway through summarizing some of Joe’s notebook notes (years '72 and '73). There is a link halfway down in the page mentioned earlier.
or here
NOTE: I’ve separated out Joe’s diagram/schematic work and all-other-discussion related notes. There are so many diagrams, I think they deserve their own summary.
The full set of PDFs has been sorted and cropped down to more digestible format. Includes some of Paul’s and others letters and notes:
voidstar78/SCAMP: Special Computer APL Machine Portable Archive Material (github.com)
Lots of great insights! Have to work on some other things - might be awhile before the next update (that would cover '74 - '75 and more of Paul’s program-management type notes)
2 Likes
Is there a SCAMP emulator available? I understand that there has been some recent work on the 5100 emulator with the inclusion of SpaceWar.
Oh, there won’t ever be a SCAMP emulator. To clarify, there are (at least) 3 terms related to SCAMP. There was some 1960s IBM SCAMP project (I don’t recall the full context of that system; or may have just been a kind of early disk drive concept?), then the 1973 SCAMP prototype to the IBM 5100, and a later SC/MP kit commonly called SCAMP.
From all I’ve read of the 1973 IBM SCAMP prototype project, only one was ever made and it resides at the Smithsonian in the Washington (hmm, which I’m not sure is open during a shutdown?). However, there is some rumors/suggestion that as many as 6 were made (some of those may have just been wooden case concepts, i.e. mockups and not full computing systems; personally, I do think there were two functional prototypes, but no one has fessed up on where that 2nd full system ended up).
The early 1973 SCAMP did use the same PALM as the IBM 5100 - however, it was a slightly earlier version of the PALM and lacked one of the shift opcodes (I forget which one - like rotate left vs rotate right). This is known since Joe George writes about it, and reports that they did a hardware workaround to add the opcode (fascinating to think that if a machine doesn’t do what you need, you can hardwire it a custom opcode! – in college in the 90s, I read about “dynamic instruction sets” and I wonder if in the future, AI-assisted software compiles will craft a custom instruction set to better optimize execution speed, but I disgress)..
While the Smithsonian does have a box of SCAMP tapes (about 30 of them), I don’t think there is any SCAMP software “in the wild.” Some things the 1973 SCAMP did was on startup, it asked a series of “what do you want to do today?” type questions. At least, there are screenshots to this effect (but I’m not sure if it was a pre-loaded APL program - since APL was the only thing SCAMP supported - the “A” in SCAMP meant APL). I don’t believe anyone has any copy of any of the 1973 SCAMP ROM software. When I was given a chance to see the SCAMP at the Smithsonian, it did have the REGISTER switch and external SINGLE systep switch (like the later 5100 production system), but I wasn’t permitted to power on the system and see if it had the special “show ROM content” during the first cycle of operation.
The IBM 5100 could have been released in 1974 - and personally, I believe some APL-only "pre-orders: of the 5100 were made and delivered in 1974 (there were some customers very eager to get such a portable system, and they didn’t care if the case was completed). These would be high end scientific shops or certain operatives, not CES-like early-look people.
But it’s speculation, the official IBM 5100 release date is in late 1975. The APL-only version was technically ready in 1974, but due to that looming 1975 anti-trust trial coming up for IBM, there speculation that IBM was obligated to show that the IBM 5100 was not “vendor locked” - to do this, they showed that the IBM 5100 could support additional “language cards” and there was a concept to also add a FORTRAN card (but that never materialized). In any case, it cost them a year to do the technical work of figuring out how to also emulate an existing BASIC. And personally, I feel these kinds of moves did help IBM years later (when in 1981 the anti-trust case was dropped).
[ many tend to forget that was a decade long anti-trust lawsuit against IBM – it also that wasn’t their first time about it; it’s a tough situation since you want to protect the integrity and reputation of your product only provide certified-good software, but you’re not suppose to sell a product that is then dependent on something else in order to even use it – these days the government seems to have given up enforcing anti-trust laws, finding it more profitable to instead just buy stocks in the company that couple service to their widgets; that said, I hear the US government is now finally mandating that all the power tools vendors agree and standardize on their dang lithium battery format, and anti-trust laws might similarly be helping to encourage auto shops to standardize on EV car charging ]
Anyhow, I’d say the IBM 5100 emulator is close enough to the SCAMP experience (note the APL used in 1973 SCAMP was derived from a different system than the APL that ended up used in the 5100 – I’m forgetting the details right now, but I think the 1973 SCAMP APL was derived from the IBM 1130 system, while the released IBM 5100 APL was derived from the S/360 {the 5110 APL was “upgraded”/expanded and derived from the S/370; BASIC was from the S/3. I think the main reason they changed the APL source for the IBM 5100 was that the S/360 had a higher precision floating point support.
NOTE: If you meant the later 70s SC/MP, I’ve not looked if there is a emulator of that.