Back To Home - Archive.org post
With the AI craze of today, you may occasionally hear references to "Expert Systems". These were, depending on who you ask, the first true software implementation of artificial intelligence. They were deterministic, working off of a large body of if/then logic to arrive at a conclusion as if they were a human expert on whatever subject in question. Interest in these systems quickly waned when artificial neural networks showed up, but until that point they were the absolute hotness if you needed something to solve this kind of problem. They first started appearing in the mid-1960s, but they exploded in the 80s with the microcomputer revolution like so many other things. Suddenly you didn't need an entire rack of computers to run one of these, and by the end of the decade they were readily available on consumer workstations and even home systems if you really wanted.
The way these worked in general is you would define a set of rules relating variables to each other, and then ask the expert system to solve for a specific variable. It would then work out what inputs to ask for to narrow down the potential outputs, before eventually arriving at whatever conclusion based on the inputs provided. If that sounds like a pretty obvious concept, well, it is. At least now. You'll see this kind of thing all over the place in software today. If you've ever used one of Microsoft's "troubleshooter" tools, like the network troubleshooter that can sometimes even work if the fix is something simple like "renew DHCP lease" or "reset network adapter", that's building on the same basic idea. Most implementations these days however have a specific sequence they check their rules in, rather than determining that dynamically. Depending on who you ask, this either means that expert systems are extinct, or inescapable.
But let's go back to the 80s for the heyday of these systems. Among the companies producing these software systems was Texas Instruments, who you have probably heard of. Most likely, in the context of retro computing, you know them for the TI-99/4A home computer, or maybe their TI-9x0 series of minicomputers. But they also made software. Among that software was an expert systems package named "Personal Consultant", which has to have been chosen specifically because of the confusing acronym because I can't imagine that nobody thought of "running PC on my PC". That has to have been the whole idea. There's not much about Personal Consultant out there, but there's not much about anything in this field out there since mostly this software was sold business-to-business, so that's not surprising. But you probably know how this goes: I want to change that.
Personal Consultant was sold in two forms, "PC Easy" and "PC Plus". PC Plus was evidently an evolution from an earlier unsuffixed "Personal Consultant", but I have nothing about that specific version. PC Plus would have run you a cool $2,950 in 1986 (or nearly $9,000 in 2026). PC Easy was much more reasonable at only $500 in 1987 (or ~$1500 today). The sales brochures also reference something called "Explorer", which appears to refer to both a TI-made Lisp machine as well as some kind of software package; I don't know if these were sold as a combination or if I'm just being mislead by something here. Either way, while these prices are expensive to modern consumers, they're well within what businesses would expect to be paying for professional software. Today, if anything, they're cheap.
Seeing a Lisp machine isn't particularly surprising, though, as Lisp and Prolog were the preferred languages for creating expert systems. I don't really know why, but I assume it's a combination of language features and just work starting there making it easier to build off existing efforts like how even into the 2000s a lot of math code was still built in FORTRAN because it had been proven there (this has mostly been displaced by Python now). TI had its own version of Lisp, too, or rather the Lisp dialect Scheme. This was creatively named "PC Scheme", but whether that's "Personal Computer" or "Personal Consultant" I have no idea. Version 3 appears to be floating around the larger internet with earlier versions unknown, and somewhere after that it was first released to the public in a less capable version and then sold on to some other company.
Onto what I actually have to show, starting with PC Scheme 2.0. It's notable to have this version, but as for the software itself there's only so much I can do to make a compiler interesting. So aside from the historical significance, I will call out two notable features. First, it came with three versions, named PCS, PCSEXP, and PCSEXT, for conventional, expanded, and extended memory respectively, but each with their own 96-98KB executable which was a fair amount of space in 1986. Second, the bulk of the application is itself programmed in Scheme which I understand is fairly common for Lisps, but in this case the files are this weird psuedo-bytecode format rather than compiled all the way to machine code. It seems theoretically possible to reverse this to source code, though I don't know enough about Scheme to do so myself.
Next, we have PC Easy v1.0. I am not an expert in expert systems, but thankfully it comes with some sample data to play with so I don't need to figure out how to build one from scratch. Each set of data is called a Knowledge Base, and included with the copy I got are PARKTRIP and MMFUNDS (for "money market"). Let's look at PARKTRIP because I think that's the more interesting concept.
Within each Knowledge Base, you can CONSULT (ask questions) or DEVELOP (modify data/rules). If we CONSULT, it asks for your name and which region of the country you want to visit. Based on that region it asks several followup questions. For example, picking SOUTHWEST asks if you want to stay overnight; answering "Yes" makes it immediately tell you to visit the Grand Canyon, and to beware of bears. If you don't want to stay overnight, it starts asking if you're interested in CAVES, FOSSILS, BATS, HISTORIC, MOUNTAIN-CLIMBING, BOATING, or WHITE-WATER-CANOEING. I tried BATS but it told me it didn't know what to recommend. In fact, no other option from that list seems to work. If you instead try SOUTHEAST, you get another list of activities like BOATING and ALLIGATORS. As a person of taste I picked ALLIGATORS. It then asks if you have any of several health conditions. I picked all of them to see what it did. It told me to go to the Everglades, but also that it's full of insects and to bring repellent. Honestly, for what I assume is a toy example, this is pretty neat if you ignore the southwest seemingly being broken.
(screenshots for PC Easy and PC Plus captured from 86box running a 386 with EGA; PC SCheme also runs fine there but I can't think of what screenshots one would want from a compiler)
Onto DEVELOP. From here you can edit PROPERTIES, PARAMETERS, and KB-RULES. PROPERTIES seems to be KB-global settings like the welcome prompt and that it needs to ask your NAME and it should be solving for PARK. PARAMETERS defines things like the four areas and what the options for health issues are. KB-RULES then starts bringing it together, with the first rule being IF AREA = SOUTHWEST AND OVERNIGHT THEN PARK = "GRAND CANYON". Based on these rules it looks like wanting BATS should have given me Carlsbad Caverns but I don't know why that didn't work. Maybe there's a typo somewhere. The point, though, is that at no point do you actually tell it what to ask the user other than the initial NAME prompt. The software internally figures out what's the most useful question to ask to refine its potential answers and presents that prompt automatically. Even the opening "which region do you want to visit" is something it figures out it needs to ask.
That's about it for PC Easy. PC Plus, meanwhile, is its bigger brother. I had to copy the contents of PC Scheme into its folder to get it to run as it first couldn't find PCS.EXE, then COMPILER.APP, and then just threw a bunch of nondescript system errors until I copied all the FSL files over. With that done, the user experience during a "consultation" is pretty similar, but this one can display graphics when needed. It's also built around "FRAME"s, with each frame having the properties/parameters/rules that the overall KB had in PC Easy, along with adding META-RULES, VARIABLES, FUNCTIONS, and TEXTAGS. You also have an option alongside CONSULT and DEVELOP to BUILD the system. Choosing BUILD generated a .RB file, though without the manual I have no idea what that's for. I also don't know what the FRAMEs are for, or why you'd want more than one, but my guess is this is how you ask it different questions since PC Easy doesn't let you do that. However, I don't really want to spend the time learning how to actually develop an expert system just to show off this thing, and PC Easy demonstrates the concept easily enough, so here's a couple screenshots just to show what PC Plus looks like in comparison including some of its more powerful features.
The ad for PC Plus states that it's compatible with both IBM and TI PCs. Did you know TI made an IBM 5150 (the original PC) competitor? Note that I didn't say "compatible". Let's go back in time a few years. In early 1983 they released the Texas Instruments Professional Computer, which is generally "inspired by" the IBM PC, but with a number of key differences. The TIPC, and its Portable cousin featuring the same hardware in a luggable form factor, have some pretty compelling advantages over the IBM PC:
- The CPU runs at a full 5MHz rather than 4.77 (which Wikipedia and others will rush to claim as a key difference, despite the others being more significant in my opinion)
- It could assign up to 768KB of RAM to the host OS, vs the PC's 640KB
- There was a dedicated RAM expansion slot, effectively giving it 6 expansion slots vs the PC's 5, though it was in-line with the card slot closest to the side of the case which restricted which cards could go there
- The floppy controller and a parallel port were on the motherboard, freeing up one or two more expansion slots (the PC's monochrome card had a prallel port on it, so if you weren't using color, you already had that one for "for free")
- It could do monochrome video at 720x300, or with a second board sandwiched to the video card it could do 8 colors at that same resolution, vs CGA's 640x200 monochrome mode or the more common 320x200 4-color modem
- A hard drive controller, an internal modem, and even a network controller were available as factory-installed options
- There was also a quasi-built-in speech synthesis system, presumably derived from the TI-99/4A's
In practice the TIPC's video modes look far better to me than CGA. While CGA had a 4th intensity bit that the TIPC did not have, giving it a wider total palette, CGA was restricted to only one of several predefined 4-color palettes at a time unless you used the very uncommon 160x100 mode. The enhanced resolution on the TIPC just looks so much nicer for text mode than CGA, even 80-column CGA. It's pretty close to EGA (720x300 vs 640x350), but that wasn't out yet. It is slightly worse than the pixel resolution of IBM's MDA card (720x350) for the same character count in text mode (80x25), but MDA can only do monochrome.
... at least, officially. While looking that up I learned that apparently very early versions of the IBM MDA display adapter actually had the hardware to drive the RGB pins on the adapter that were otherwise unused, but they became physically disconnected later on. Maybe TI was inspired by that, but even if not I think this display adapter is a pretty good "best of both worlds" for when it came out.
The TIPC sure physically looked like an XT too, except TI made up these neat faceplates to do half-height drives vertically centered in the full-height bays, which I think looks way better. Of course, early models were still using full-height drives, so whether any given TIPC has those is a coin toss.
It's also got a number of other weird evolutionary differences from the PC as well. There's a back panel connector for a 3rd and 4th floppy and in fact the hard drive mounts at E: rather than C: because it assumes you'll want to use them. Its keyboard uses RS232, but runs the transmit and receive lines at different baud rates, which it turns out is a thing you can do with the 8251 UART. It did drop the cassette port, which I'm actually kind of sad about because it's a change that makes it more normal instead of less. And they used a bunch of hex-drive-only screws, so I had to get out my socket wrench to take it apart to check for RIFAs (which it had, and also four DC-side filter caps that absolutely reeked and were starting to corrode the power traces; if you own or buy one of these check that!). My TIPC did come with a hard drive, but despite a herculean effort on my part I cannot get it to cooperate (I reinstalled it for the photo for the visual). I think the read heads are fried. Note that the floppies I'm sourcing this post from were acquired separately, so I don't have reason to suspect there's anything particularly valuable on the drive, I'm just annoyed it doesn't work.
TI ran a number of ads highlighting the key differences, most under the "Dare to Compare" tagline. These advertised faster speeds, better expandability, better colors and screen resolution (compared to CGA), built-in support for 320KB floppies vs the PC's 160KB, and even how the TIPC keyboard had separate arrow keys and had 12 F-keys vs. the Model F's 10, both features that would become standard for IBM but hadn't yet. Several ads also comment that the keyboard has "isolated edit/delete keys", which I also assume is in comparison to the IBM Model F but I've never used either in person so I can't speak to how much of an advantage that was. There was also a decent focus on the TIPC having onboard printer and floppy controllers to show how the TIPC had more flexible expansion options, but IBM's solution was to release the XT that added 3 more slots which arguably came out ahead in the end.
In exchange, the memory layout was completely different, the expansion card pinout was subtly changed from the IBM PC (and it didn't support DMA), and there were a handful of other differences along the way. This meant that while software written purely against DOS calls could run natively on either the IBM PC or the TIPC, anything that used the hardware directly would only work on one or the other. And that was was most software, since you needed to do that for basically any software to run well. Nor would most expansion cards work in one or the other due to the different memory layout and the changes to the slot pinout. So while these enhancements are nothing to sneeze at, you still need to convince people to switch and leave their existing software libraries and hardware suppliers behind.
They didn't, of course. TI enjoyed reasonable sales of the TIPC, but then the AT happened and the PC clone market exploded and the TIPC was left in the dust. It lingered on for a while before being discontinued somewhere around 1985, though I can't find an exact date.
Not much of the TIPC's software library has survived. There's a pretty comprehensive archive on, well, Archive, which includes titles like MS-DOS and CP/M-86, pfs:file and pfs:report, EasyMailer II/EasySpeller II/EasyWriter II, an IBM PC emulator, Microsoft Flight Simulator, and my old friend, GSX-86. There should have been a TIPC version of UCSD p-System in addition to MS-DOS and CP/M-86, but it seems likely that has not survived. Another TI ad also lists Multiplan, Lotus 1-2-3, Wordstar, The Home Accountant, and dBASE II, but those also seem lost.
All of that is pretty reasonable software for a professional system, but all of it except things like the TIPC diagnostics disks are third party programs. MS-DOS and Flight Sim are from Microsoft, the IBM PC emulator is from Romberg & Romberg (what a great name), pfs:write and pfs:report are from Software Publishing Corporation, and so on. And all of those were things you could already get on the IBM PC, even if they'd run and look better on the TIPC. I don't have any of the missing commercial software, like Multiplan. What I have is more exciting: software by TI, for TI, never intended for commercial release.
Texas Instruments was and still is a large company, of course. Like any large company, it equally was and still is comprised of many different divisions. The TIPC team itself, referred to as "TIPC" internally as well, was one such division. But they were a hardware division, not software. Instead, this incredibly long segue is so I can talk about the other software I got, which came from TI's Integrated Software Support Systems (ISSS) division. ISSS did internal software development, and as such they had problems they had to solve to keep TI running well as a business. With the addition of the TIPC they now had new capabilities, and they set out to take advantage of those.
Texas Instruments had already defined itself in the world of computers through its minicomputer line, with the TI-99/4 and /4A (in)famously being derived from the TI-990's architecture. Internally, TI had multiple such systems that users often had to connect to. Previously, they did so using a dumb terminal like a VT-100. Why not do that from a TIPC instead? Then, you could disconnect from the company mainframe and work on your sales spreadsheet without having to switch computers. So ISSS started work on an emulator for the VT-100 (it also names the TV-970, but that appears to just be a terminal compatible with the VT100 among others; I assume TI owned and used both).
The VT100 emulator (and everything else in this post from here on) requires a TIPC to run. To my knowledge there's no emulator support for this system anywhere, unless it's buried somewhere in MAME. Screenshots for this and the following entries were captured from my TIPC using the RGB2HDMI capture function. I think I have the phase off slightly so if the edges look slightly weird that's my fault, I'm not an expert at that part yet.
It's... not much to look at, though. It's a terminal emulator, after all. I tried to get this to actually do its job as a terminal, but it didn't want to cooperate with my TIPC's serial card. I don't know if it's because my card is third party, if something's wrong with the card, or if I'm just using the software wrong. But I have several copies of this through its development versioned 0.3, 0.4, and 1.0, including the source code in C for some versions. Included in that source, which is fascinating to perhaps literally only me, is this from "VT100REV.DOC" from a disk labelled as having file transfer capability:
On March 21st, 1985, a meeting was held between LAN Software Development Group (LSD) and [name] of Engineering and Manufacturing Department (EMAD). The main issues arised were the unsatisfactory binary file transfer speed and various file types and file organizations under Apollo. We reached an final agreement that our group (LSD) is committed to continuous support of VT100XFR products, we would investigate the feasibility and possibility of improving binary file transfer benchmark.
The old binary file encoding scheme were very straight forward, that is, split one byte of binary data into two bytes of encoded data such that transmission over NIU, modem or other medium would be fool-proof. However, elapsed time of a binary file is twice as long as an equal-sized text file. Let's call this scheme 'Two for One'.
We have investigated the possibility of employing 'Hamming Code' algorithm. We feel this is an over-kill for slow speed file transfer program. CPU data crunching time out-weights the saving gained on better data compression scheme.
The binary file encoding scheme selected is named 'Four for Three', that is, to encode three bytes of binary data into four bytes of 'safe ASCII' characters. This method result in 33% reduction in elapsed time comparing to old scheme. Also comparing with text file transfer, the 'Four for Three' scheme is only 33% slower instead of twice as slow.
I believe based on some other documents that "Apollo" in this context is the Apollo line of workstation computers. The really neat (to me) thing is this is effectively a several years early equivalent of what Base64 would eventually do. This kind of documentation rarely survives, so it's fun to see it in the wild. But it's also noteworthy in that it's a file transfer utility for a terminal emulator, and that's less common. I also believe they're using "LAN" here to refer to serial connections as well as what we'd think of as "network" today.
Building on the VT100 emulator, ISSS also authored a piece of software called PC-NET. The bad news is that like the VT100 emulator, I have no idea how to actually make this do anything, possibly due to the same serial issues:
The good news is it comes with a comprehensive manual so we can at least learn how you would use it if it worked. PC-NET is a distributed network file editing system, from 1985. It does seem to need a host component that I don't have, so even if I could figure out the serial card I probably can't do much with it. But the intended workflow is that you'd log into a host (mini)computer and then from there pull up specific files and edit them from your TIPC. It's even designed to let you have two documents open at the same time, taking advantage of the TIPC's larger screen resolution. I have versions 1.0, 2.1, and 2.11, though only 1.0 has the manual. I also have "OSE" source for 2.11. It's not clear to me what OSE actually is, because it's all described as just things like "This procedure sets the current pointer to the edit_block which contains the requested index, or return END (of file ).". It's possible this is a part of the server-side component, but whatever it is it is definitely incomplete.
Naturally, if you try to Emulate from the menu, our friend the VT100 emulator returns:
Next, we have two more terminal emulators, for a 914 and 931. They are what they sound like, but there's a couple things worth talking about. The 914 emulator includes VCHK.EXE, which appears to yet again be related to transferring files. This time though it's specifically used for synchronizing certain local files needed for certain IMS transactions with the host system. IMS here is IBM's Information Management System, a mainframe database system which we'll come back to in a moment. I don't remotely have a setup to test this, and I don't see any filenames in VCHK.EXE so it must get that list from the server so I can't even speculate as to what it was used for. The 931 emulator meanwhile has a manual for something called "IBM/TI PC Directory File Editor (DF-EDIT)", which sadly I don't have. It's also got a couple files that imply that it was somehow using the TIPC's speech synthesis for something, but I don't think this is complete, either.
Next up is Business System Access, which is a frontend for some kind of internal TI mainframe service. This is the driest thing imaginable and I absolute love it. It also doesn't work, throwing an error on startup; it might need either the Ethernet card or the modem card as I have neither. Or maybe it's just that third-party serial card misbehaving again. It's not like there'd be much to show if it did, anyway, since I don't have a TI mainframe for it to dial into much less whatever specific host software this was meant to be used with. Weirdly there appears to be two different versions in the same folder, with v2.1 being EXE files and v2.2 being COM files. Each has executables for BSA, BSADRIV, DISP, INITSDB, KBDSP, KEYBRD, MUSTART, NAMEMGR, NETSTAT, NIOS, PRNT, SHFTPRNT, XNETP, and XNSECHO, with just two standalone files of BSACFG.EXE and NIOPDRVR.DEV. We'll all have to collectively guess what those all did, though the names are somewhat obvious for some of them.
... and now attaching these images I see that the splash for that .COM says 2.1. Maybe 2.2 was in development and not done yet?
Then we have two training courses for IMS, as above. I assume TI must have been investing heavily in IMS around this time. At least unlike Courseware, for this one I have the runtime so I and the tens of other people with TIPCs can experience it. It seems pretty good for what it is, and it's another decent demo for the TIPC's text rendering:
For the last two entries here, references to "TIPC" disappear and are replaced with "PWS". What's "PWS"? I'm honestly not sure, but I suspect it's the group that took over when the TIPC was discontinued, if not just a renaming of the TIPC group itself. "Professional WorkStation", maybe. The TIPC wasn't in fact TI's last homegrown x86 desktop system. That would be the improbably terribly named "Texas Instruments BUSINESS-PRO(tm) Professional Computer". The BUSINESS-PRO notably had 14(!) ISA slots, but it's substantially rarer today than even the original TIPC. A random VCFed post says that the BUSINESS-PRO was a true IBM compatible, but that's all I've managed to dig up on that front. It's hard to prove that this thing even existed aside from a manual on Archive and somebody listing a badly battery damaged one on eBay.
This being a true PC compatible would line up with Personal Consultant running on regular PC compatibles. You could buy a TIPC-compatible video card for the BUSINESS-PRO, allowing you to keep your old software library as you upgraded to TI's new system. In theory, anyway, since you'd still have to deal with the different memory map and all the other changes. Maybe they had a reverse emulator, since I found at least one ad for the BUSINESS-PRO that claimed it could run TI's entire software library.
Still, I've got two more programs to show off here. Both of these really want to run from a hard drive, as above I cannot get my TIPC's hard drive to work. They'll kind of run off a floppy, but some functionality isn't available as they depend on launching other exe files for some functions. I did my best.
PC Shell is, well, a DOS shell. Nothing seems to say if this is "Personal Computer" or "Personal Consultant", but I can't find any reference to it otherwise so I'd bet the former for this one. If you've ever used the DOS shell that comes with MS-DOS 5, you've got the basic idea, but it also has a side by side to-do list because why not. It makes pretty good use of the higher resolution display, but otherwise is unremarkable, being one in a very long line of these things to be created in the aftermath of Apple demonstrating the future of computer interfaces to the world while also completely not understanding why the Macintosh was so popular. Nothing I see here implies I should be using a mouse so they don't get credit for that either, though no ad I saw for the TIPC showed a mouse... but I believe it did have one. More on that towards the end.
Professional File Manager on the other hand is a file manager that can also show you upcoming appointments, which I have to admit is a combination I've never seen before, probably because I can't imagine a situation where I'd be spending so much time managing files that I'd need to know that it. But I suppose in the 80s this was a believable future for computer users. It also has some other functions built into secondary executables that it can launch, including the ability to "suspend" by launching a DOS prompt from inside PFM, but due to the aforementioned hard drive issues I can't really demonstrate them. It's otherwise a pretty capable file manager and once again, this looks great compared to what the IBM PC could do at the time:
You can even change the color theming. Which I did, because as I found it, it looked like this:
I have absolutely no idea why anyone thought this was a good idea. Maybe it was just an example of how far you can customize it. Either way, my eyes were unhappy. I would generously describe this color configuration as "difficult to use" and less generously as "poop from a butt". To be slightly nice to it this specific screenshot is from before a friend pointed out to me that I had the wrong palette loaded in the RGB2HDMI so the oranges should actually be yellow, but I don't think that would improve it much.
While PC-NET and the emulators appear to be internal, and PC Scheme/PC Easy/PC Plus all had commercial releases, I have no idea about PC Shell or PFM. They'd be odd programs to keep internally, but PC Shell is from 1986 and specifically branded with ISSS, so it seems unlikely it would have been sold for the already discontinued TIPC. Maybe it worked with the BUSINESS-PRO, but it'd still be odd to sell it with that branding. PFM seems more likely to have been a proper product, but I have no evidence either way except that it too doesn't run on a regular PC.
One last thing to cover, then. In the process of trying to get my TIPC to cooperate I picked up a copy of TIPC Diagnostics v2.1, which apparently does not exist in the public as all anyone has online is v1.2/1.4/1.6. Oddly v1.6 has speech tests that v2.1 doesn't, but v2.1 adds a bunch of new tests: "EMU", a 3270 connection kit; LAN, for the ethernet controller; Mouse (so it did have one, though I have no idea which); RTC, which I don't believe was ever made by TI but I guess was enough of a standard to show up here; and "S100 Transport" which I am now incredibly sad that I don't have. The disk is dated 1984, so this wasn't even that far into the computer's lifespan to have all these options available.
So the grand total of what I have is this:
- The Diagnostics v2.1 disk, which is a proper commercial disk with TI label and everything
- Backups of development versions of the VT100 emulator (0.3, 0.4, and 1.0), and PCNET (1.0, 2.1, and 2.11), on consumer disks
- A disk named "CABLES - CABLES" that contains diagrams of RS232 pin mappings for various systems, which I'm including in case it helps somebody somehow
- A 12-disk PC backup, which includes the 914 and 931 emulators, a different version of the VT100 emulator, BSA, PC Shell, PFM, PC Scheme, PC Easy, PC Plus, and the usual DOS system components
All of this has been uploaded to Archive. As far as I can tell only PC Scheme, PC Easy, and PC Plus can run on regular PCs. On the other hand, those are probably the most interesting things here anyway. If you want to try those, pretty much any XT- or AT-compatible or better machine should work if it has enough RAM, though it might need CGA at minimum, I never tried MDA or Hercules. I ran mine in 86box on a 386 with EGA and 640KB of RAM, but the TIPC itself is just a 5MHz 8088 so even an original PC should work. If you actually have a TIPC, then you just need a hard drive to run PC Scheme/PC Easy/PC Plus, and PC Shell and PFM want one but will technically run without it, but the rest should run from a floppy fine. At least, as well as they run in the first place.
All the above have been uploaded as both .img as well as .raw flux files. Note that disk 2 for the IMS training course has a few bad sectors, but those sectors appear to have been blank so I don't actually think it matters. Some of the other disks required me to clean some fairly aggressive mold off to get good reads, but this one seems fine in that regard. I tried several times in different drives and got the same result every time so I think those sectors are just legitimately bad.
There were a couple other disks in this set (the CourseWare disks were actually in this group too), but they were all commercial non-TIPC software that's been well preserved like dbIII and PKZIP, plus a couple personal artifacts that are of no historical value. Those have not been included.
Hopefully you enjoyed this somewhat all over the place post, since there wasn't really a great way to tie all of this into one single narrative. But PC Easy and PC Plus are absolutely notable enough to be worth mentioning, and the rest are still worth preserving if just for the novelty of it.
And just remember, if you ever visit the Grand Canyon overnight, beware of bears.