Advent of Computing - Episode 63 - What's With The Wedge, Part 2

Episode Date: August 22, 2021

This concludes my series on the distinctive shape of early home computers. In this episode we finally cover the Sol-20 itself, the first system on the market to be shaped like a wedge. More generally, we try to figure out if the Sol-20 was the progenitor of hundreds of machines that followed, or if the wedge was inevitable. For such a simple question, this has become a surprisingly complicated topic. Selected sources: http://archive.computerhistory.org/resources/access/text/2012/10/102702231-05-01-acc.pdf - Lee Felsenstein, oral history at CHM http://www.leefelsenstein.com/wp-content/uploads/2013/01/I_Designed_the_Sol.pdf - Article about the Sol-20's design process http://www.leefelsenstein.com/wp-content/uploads/2013/02/Felsenstein-Tabloid-BW.pdf - Tom Swift Lives! Like the show? Then why not head over and support me on Patreon. Perks include early access to future episodes, and bonus content: https://www.patreon.com/adventofcomputing

At the beginning of this month, I had a simple question. Why were so many home computers shaped like wedges? Seems easy, right? That shouldn't take too long to figure out, and it should make for a simple, fun episode. Just find the first angular machine, figure out why it took on that distinctive form, and then look at how the idea spread to the next few computers. Hey, that's even in three acts already. It basically writes itself.

Well, dear listener, nothing can ever be that simple, especially when it comes to the complex history of computers. The more I looked into this question, the more my net expanded. It's not just the shape alone. Home computers of this era, from the latter half of the 70s up to the middle of the 80s, tended to follow a really rigid formula. You got a molded plastic case with an inclined face for a keyboard. Inside was a motherboard with a microprocessor and some RAM thrown in. Around the back of the case, you have outputs for TV signals and some kind of expansion slots.

These could come in different flavors, ranging from simple ROM cartridges all the way up to more complicated buses. That's the general blueprint, and a surprising number of computers all fall under the same umbrella. You get your Apple IIs, the first model of the TRS-80, RadioShack's Color Computers, Dragon 32s, and Spectrums. Basically, every computer that Commodore ever made, with the exception of the PET and some later model machines, also fit this plan. That ranges from VIC-20s all the way up to the Amiga. Atari 400s, 800s, and STs also adhere to this grand plan. Sinclair, Acorn, and countless other manufacturers from around the world all fell under the spell of the wedge. We also know the first computer on the market that met all these criteria. That's 1976's Sol20 terminal computer.

It was designed by Lee Felsenstein for processor technology. So this should be one of those things that I can just button together, right? But there's this bigger question. How could all these ladder machines be so heavily influenced by the Sol 20? Were they all just ripoffs of this one computer? Or is there something a lot deeper going on here? Welcome back to Advent of Computing. I'm your host, Sean Haas, and this is episode 63, What's with the Wedge, part two. And yeah, as you can tell, we're back into this weird rabbit hole I've dug for myself. At the start of this month, I decided to tackle the whole wedge question around early microcomputers.

And to do that, I figured the best starting point was to trace the story of the first wedge-ish machine. That's the Sol 20. Last episode was mostly the preamble to the actual story, covering all the background and building up a web of context. If you haven't listened to that episode, then you probably should. The content within will be really crucial moving forward. This episode, we're getting into the Sol 20 computer proper. We're going to be looking at its development, its features, what's inside it, and how it would impact the industry at large. And yes, we will try to explain why it turned out with that distinctive slant in the front. Before we do that, I want to kick things off on the same note last episode ended on.

We're deep in the story of early home computing, but it's important for us to keep in mind that the rise of personal computing had been really slowly burning for decades. that the rise of personal computing had been really slowly burning for decades. It started even before the first microprocessors. Depending on how you really look at it, folk were thinking about personal computers before the first mainframes started crunching numbers. So this is a pretty old quest. What makes this era we're exploring so unique,

this period in the middle of the 1970s, is that it was the first time that outsiders were able to enter the computing industry en masse. On this podcast, I've spent a lot of time talking about how people with new perspectives can be uniquely qualified to solve problems. The most recent example on this show is Ted Nelson. He wasn't a computer scientist, and he definitely wasn't a programmer, but he was able to push hypertext from very rough theory into something much more tangible. Outsiders are an important force against stagnation. With the release of the Altair 8800, the creation of the Homebrew Computer Club, and all the splinter groups and companies that follow, we can see a flood of outsiders.

we can see a flood of outsiders. Armed with microprocessors, coordinated with clubs and newsletters, these outsiders gave computers the final push they needed to enter the home. But that wouldn't have been possible without earlier innovation and earlier breakthroughs. You don't get an Altair 8800 without an Intel lurking behind it somewhere. The point is, we're entering back into some complicated territory. I'm going to stop rambling, because it's about time we get back to last episode's story. At the end of last episode, it was November 1975. Les Solomon, editor of Popular Electronics, had just asked Bob Marsh if he could make an intelligent terminal for him. At the time, Marsh was running an outfit called Processor Technologies,

sometimes shortened to ProcTech. And I'm going to be shortening it to ProcTech because that's a lot quicker to say. Solomon's offer was pretty simple. If ProcTech could build a prototype to show off, and eventually a fully working model, Popular Electronics would run that machine as a cover story. According to legend, the proposed timeline was one month, but that wasn't really ever stuck to. Marsh called up the one person at Proctec that was well-suited to the job, one Lee Felsenstein.

At least, Felsenstein had the most applicable recent experience. He had just finished designing the VDM-1, the Video Display Module. He had just finished designing the VDM-1, the Video Display Module. This was, for lack of a better term, a graphics card for the Altair 8800 computer. It provided a way to get text onto a TV screen, something that was impossible to do with a stock Altair. Initially, users needed to get a separate terminal to get much use out of an Altair, but with the release of the VDM-1, that changed. Add a keyboard interface, a keyboard and a TV, and boom, you all of a sudden turned your Altair into a standalone system.

Also, a big step along the way towards fulfilling Felsenstein's long-sought goal of a hobbyist-built terminal. Since his days back at Community Memory, Felsenstein had been trying to find or build a cheap, reliable, and easily repaired computer terminal. When he started looking, there just wasn't anything that fit those criteria. In 1974, he worked up a theoretical design for his dream machine. He called it the Tom Swift Terminal. By the latter part of 1979, a tricked out Altair with a VDM-1 got pretty close to this dream. You just needed to add a serial interface and a modem, and you had yourself a pretty reliable intelligent terminal. So when that day in November rolled around, and Marsh asked if Felsenstein would be able to make a terminal on a tight deadline, the answer was pretty easy.

This was something Lee had been working towards for years at this point. Most of the pieces were already in place. This would give him the chance to bring everything together to form his exact vision. This would be a group effort, there would be some compromises, but the project seemed both possible and really exciting. One huge help was that Felsenstein had most of the design work already in the bag. Last episode, we spent a good deal of time talking about the Tom Swift terminal. Well, Felsenstein didn't just leave that idea behind in 1974. It stuck around and continued to see changes.

Sometime in 1975, I don't know the exact date, sadly, he published an article on his terminal design in the People's Computer Company newsletter. It was titled Tom Swift Lives. It's a lot more conversational than Felsenstein's previous techno-manifesto. The article describes a terminal built around a 44-pin bus, cards slotted in to handle TV output, keyboard input, serial communications, and so on. And optionally, a microprocessor could be used to turn the Tom Swift terminal into a fully-fledged computer. But there are two details that really stick out to me that were lacking from the 1974 paper. The first might seem a little silly. It's right in the article's title. Next to Tom Swift Lives is this passage, quote,

Our motto, if work becomes play, then tools must become toys, end quote. On the surface, that just sounds like a fun little aphorism, but I think this is crucial to how Felsenstein viewed computers. This also really speaks to the shift that made home computing possible. Felsenstein looked at computers as, well, fun. Once again, that just sounds trivial, but we have to put this in context. 1975 is still an era where big machines by far outnumbered microcomputers. That is, if there are even enough microcomputers to count at this point. A computer cost huge sums of money. The cheapest system you could buy was still selling

for on par with new car kind of bucks. Just as an example, in late 1975, IBM released the 5100, their first quote-unquote personal offering. They were very much trying to get into the home and office market. The computer fit on a desk, was meant for a single user, and cost just under $9,000. Adjusted for inflation, that is $45,000 today. That's not cheap in any regard. While computers were getting smaller and more were being manufactured, there weren't easy ways to access them. Individuals usually didn't own computers, at least not how we do today. You would access a machine through your work, school, or your job,

or maybe some larger institution. In that context, it wasn't as easy to just play around on a computer. Sure, some people still did, but those are exceptions that prove the rule. If you asked someone why they had access to a computer, they probably wouldn't have said, oh, you know, just to have some fun on the weekend. Felsenstein wanted to work towards a future where someone could own a computer just for fun. Part of that was personal. He felt that machines were fun, not just to use, but there was fun to be had in the very wires and chips inside. The fun of a challenge, the fun of understanding a complex device, and just the fun of blinking lights and switches. To rephrase this in another

way, Felsenstein wanted to make computers that could be used casually. That's something that I bring up whenever I get into personal computer territory, and I think it's very applicable here. The kind of people who worked with, say, an IBM 5100 weren't very casual about it. A 5100 was going to be part of your office, used for running calculations or managing data. That's a pretty serious suit-and-tie affair. There's no room for casual fun. Cheaper, more accessible, and easier to understand computers would be needed to break out of that mold. By making computing fun, or at least casual, that mold. By making computing fun, or at least casual, more people could be brought into the digital revolution. Okay, so that's the first big takeaway from Tom Swift Lives. The second is a diagram. Right next to the title, taking up maybe a quarter or fifth of the page, is a diagram of

Felsenstein's Dream Terminal. The best way I can describe the drawing is, well, a pretty chunky looking wedge. It shows a box with an inclined top and integrated keyboard, but the back of the wedge is continued into a larger rectangular case. This back portion contains the actual guts of the machine. You get a power supply, the 44-pin bus, and a series of expansion cards for memory, I.O., and video output. For me, hot on the trail of the first major wedge, this is a watershed moment. The first printed description of the Tom Swift terminal specifies that the keyboard would be a separate unit. Then, in 1975, it changes into a single-piece construction. Keyboard and computer combined into one larger

whole. Was this the first time someone sketched a wedge-shaped keyboard? No. With the human wrist being how it is, a slant just helps out. We get inclined typing surfaces going all the way back to typewriters. But we are seeing Felsenstein approach a fully wedged computer, and more importantly to the larger story, this is the first time we have a wedge in the context of all the other bits and pieces that would make a recognizable home computer. The point is, Felsenstein had a really good idea of what he wanted to build. He had a very specific goal to build towards, and even some of the technical pieces worked out. So when Marsh asked for a new computer, Felsenstein was in a really

good position to say, sure, we can do that. But this project was still approached with some caution. Felsenstein knew it would be possible to put together a whole new computer, but that may take some extra time. The initial request from Solomon was for a terminal, and that would be by far the easier goal to achieve. Marsh, however, thought that they should go whole hog and just put an Intel 8080 inside the machine, making it a full computer. Felsenstein was shooting for the easier goal. In his words, quote, alternate designs were roughed out and prices compared. Both were for terminals using the basic circuitry of the VDM-1. His, meaning Marsh's, had an 8080 thrown in, mine had decoders and counters. Mine was cheaper, but only by about $10. Gradually, I came to realize that if the inevitable were

to happen, it would be better to be on the inside than on the outside." In other words, there just wasn't much of a reason not to upgrade to building a full computer. Felsenstein's initial design could function without a processor just fine. That was a design constraint going all the way back to the days of the Tom Swift terminal. However, that still took some kind of glue to hold everything together. You still needed some logic chips and some extra circuitry. That could all be accomplished just as well via a microprocessor and a little bit of code. This almost reads like a cheesy infomercial.

For just an additional $10, you get all this and so much more. By adding a microprocessor to the mix, Proctec's new machine would be more than just a terminal. It would be a computer, and going off early designs, it would be a more capable computer than the Altair 8800. But this led to a bit of an awkward issue. You see, Popular Electronics was kind of an Altair shop at this point. Marsh didn't want to step on anyone's feet. Plus, Solomon requested a terminal. He did want an intelligent terminal, but he hadn't asked for a computer. It's a fine line between the two types of machines, to be sure.

Marsh decided that that line would all come down to branding. So when it came time to name this new project, he went with a bit of a mouthful. The Sol 20 Terminal Computer. Not a computer, not a terminal, but a terminal computer. The Sol part was a bit of a joke that Felsenstein threw in. Around ProcTech, they had said that this new machine would be built with the quote wisdom of Solomon, as in less Solomon, hence Sol. 20 seems to have just been a cool number thrown on the end to make it sound official. That brings us up to the actual design of the Solve20.

As I keep restating, the core of the computer was its bus. Felsenstein originally wanted to use a 44-pin bus, but the final product ended up using 100 pins, just like the Altair. In fact, the Solve20's bus was almost a copy of the earlier machine's bus, but there were a few key tweaks. The small wires and tiny changes on the bus may seem like tiny, kind of stupid minutiae, but it shows the stressful design process at play behind this early computer. Felsenstein used some pretty colorful language to describe what it was like working with Marsh. He's been pretty consistent in his disdain for the overall process. The main gripes were that Marsh always had new ideas about how to do things and some grand new features that just had to be added. I can personally attest

that those qualities make for a killer combo, at least a combo that will kill a lot of productivity. combo. At least a combo that will kill a lot of productivity. Anyway, one of Marsh's early suggestions had major ramifications on the Sol20's bus. Initially, Marsh wanted the computer to support an external expansion chassis that would be connected to the motherboard via a ribbon cable. This, dear listener, is a bit of impractical design. In Marsh's mind, this would work out so that the consumer could buy a base Sol 20. Then, if they wanted to upgrade, all they had to do was buy a separate expansion chassis. This would come as a metal box that would sit underneath the main system. A thin cable around back would link the two devices together.

As fancy as that all sounds, there were some technical issues to deal with. Firstly, ribbon cables can be pretty bulky. They're just bundles of insulated wires that are all bonded together in a flat sheet. The more wires you have, the wider the cable. A 100-wire ribbon cable would be a few inches across, and they weren't really used in anything. It's a comically wide piece of wire. And since they aren't very standard, they either don't come cheap or you just can't buy them. So the initial plan was quickly changed to use two 50-pin ribbon cables. But you see, that's just one of the issues with Marsh's idea. The next problem was interference. Normally, interference is only really an issue with analog signals, but digital interference can also occur. Wires carrying digital pulses can produce a small

and rapidly changing magnetic field. Normally, that just doesn't matter. It's a really localized effect and it's really weak. All you have to do is just space wires a little bit apart, maybe shield particularly noisy lines, and problem solved. Early Altair models were somewhat notorious for failing to take this into account. The whole data in-out part of the bus was particularly troublesome. Early Altair models were somewhat notorious for failing to take this into account. The whole data-in-out part of the bus was particularly troublesome.

Interference between the data traces on the backplane could lead to some strange errors, and the main issue was they were just too close together. Marsh's expansion chassis plan was basically to move this Altair bus from a circuit board onto a ribbon cable. You don't have control of where your wires go on a ribbon cable, they're all right next to each other. Data lines would be even closer than on a faulty PCB. The risk of interference would be even higher. Luckily, Felsenstein knew what he was doing and was able to find a workaround. He dropped the whole data in, data out lines and replaced them with a single bidirectional data bus. That sounds a lot more complicated than it is. Basically, instead of having 8 pins for data heading into the CPU and 8 for data leaving, Felsenstein just set up 8 data lines for transfer in general. That freed up 8

wires on the ribbon connector, which Felsenstein placed between each data wire. Those lines were then tied to ground, essentially making little barriers between noisy lines. Now, why am I getting into this kind of really tiny minutia? There's a simple answer. The bus design is a really big deal here. Felsenstein's fix had the immediate impact of solving Marsh's expansion chassis woes. It also made for a cleaner design. Felsenstein had been disappointed in the split data and outlines on the Altair bus since he first saw the computer. A processor can't send and receive data at the same time, so it just makes better sense to have a single data bus with a line for signaling its direction. I think the story also illustrates the working relationship here.

There's a lot of back and forth going on at all levels of design. Perhaps more importantly, this also meant that the Sol 20 would never be compatible with Altair expansion cards. The external expansion chassis was eventually dropped as just being too much of a hassle, but the changes to the bus design stuck around. Of course, a little change like this can have huge effects. I think this tiny technical tweak saved the Salt 20 from living as a strange Altair clone. There were still a lot of similarities, but they were two distinctly different machines. Let's zoom out a little and start to examine some of the larger design details of the Sol.

I've been talking about how the expansion bus was central to its design. Modularity was a really key goal for Felsenstein. But not all cards on the bus are equal. The Sol 20 essentially had a motherboard, but there's some caveats to that that I want to get into. The computer's manual calls this component a, quote, single board computer. Functionality-wise, this is the bulk of the system. It's a large single board that contains an Intel 8080 microprocessor, one to2 kilobytes of RAM, circuitry for handling keyboard inputs, and all the important video output stuff. Felsenstein made a simplified version of his

VDM-1 and crammed it onto the motherboard. So, all in all, we're looking at a pretty dense board. On its own, this has more functionality than a stock Altair 8800. On paper, that sounds nice. Just put a whole computer complete with graphics and inputs onto a huge circuit board. But luckily, the real world stepped in to save ProcTech from an easy project. The worst part came down to layout and routing. That's the process of figuring out where to put components and how to connect them all up on a circuit board. You start with a schematic, which is just a drawing of how everything should be connected, and you eventually end up with a mask for a printed circuit board. This process is one of those things that's more of an art than a science.

Today, we use computers to streamline a lot of that work, but back in 1975, unless you had a mainframe, you had to do it all by hand. Felsenstein and another contractor tag-teamed the task. The process at the time was to create an enlarged negative of your circuit board using transparency sheets and tape. This had to be done over a lightbox

to make everything as visible as possible. This also introduced a fun issue. Proctec was still operating out of the garage that Marsh and Felsenstein were renting. As the shared workspace became the offices of a computer company, things started to get a little cramped. The only space large enough for a lightbox was a tiny loft in the back of the garage. The working conditions in this loft weren't exactly spectacular. From Felsenstein, quote, I personally made sure that an electrical conduit running at forehead level was padded, and tried to get someone to put up a fence to prevent one of us from stepping backwards off

the edge of the loft. That, the knocking of heads against rafters, and the sticking or scratching of various parts of the head with needlepoint exacto knives, which we constantly had to hand, were the main hazards of working there. In addition, there was a chill at night and the heat that accumulated during the day." So, there is something to add to your mental image of the birth of home computing. This two-man team worked long hours in cramped conditions, trying not to cut themselves or fall to their doom, all the while slowly building up the Sol's motherboard. Computing isn't really the most dangerous job in the world, so you gotta grab onto any high-risk stories you can find. All jokes aside, this was really grueling work, and I think it's important for us to bear

this in mind. A small team of enthusiasts can beat out huge lumbering companies, but it takes more than just a good idea. It takes a lot of effort, and it takes a lot of time. According to Felsenstein, the layout was completed by Christmas of 1975, so we aren't on the one-month timeline that we started with. That seems to have been dropped pretty quickly, but we're still moving along at a really fast pace. At this point, you might be wondering, where does the whole expansion factor into this? What about the bus? Well, this is what gets us to why I'm a little hesitant to call the Sol single board computer a quote motherboard, at least not in a sense that we'd really recognize. Smack in the middle of Felsenstein's beautiful circuit board was an expansion slot, specifically the 100-pin socket.

A riser card slotted into that. It's basically an S100 backplane with a 90-degree connector on one end. Expansion cards were then plugged into this riser. Basically, this let you slot in cards parallel to the main system board, which was convenient for packing it into a case later. Now, here's the dig. There isn't anything like a bus controller at play here. There's nothing fancy managing that bus.

We're still dealing with a simple backplane-style bus, just like on the Altair. So the Sol 20's single-board computer is, for all purposes, just a really physically large expansion card. It has all the basics you need for a working computer, right down to RAM and I.O. all on one card. But at the end of the day, it still plugs in to a 100-pin backplane. If it weren't for some changes in the pinout, I bet you could wire the same base Sol board onto an Altair 8800's bus. I think this all comes down to DNA left over from Tom Swift. That theoretical system was all about modularity and expansion over a simple bus. The Sol 20's strange motherboard-slash-super-expansion card is a fitting heir to that idea. If the guts of Proctech's new computer were a mix of Tom Swift and Altair,

then what about its exterior? Well, dear listener, this is where we finally get to the first real wedge of our story. This is also where we veer into some weird territory. Bob Marsh had this friend, Steve, who was a carpenter. That relationship would color a series of Marsh's business decisions. Steve was something of a silent partner in proc tech. He built the light table that was used to lay out the Sol's circuits, for instance. And he would also factor heavily into the Saul's physical appearance.

From Felsenstein again, quote, Steve had told him of a great bargain to be had from center-cut pieces of walnut, which were ordinarily almost thrown away. There is some cause to believe that the primary reason for existence of the saw was to provide an outlet for this inexpensive wood, end quote. It's a funny turn of events to be sure. However, we can also glean some insight from this. For starters, this should serve as another reminder that we're dealing with a very small company. You could almost say scrappy. Resources were limited, corners had to be cut, and in this case that meant literal scrap wood had to be used. I wouldn't say that the walnut

panels were just an example of Marsh having eccentric ideas. It sounds more like an attempt to save money on material costs. A company like IBM would never dream of doing something like this. But cutting corners was a way of life for ProcTech, and really for this whole first wave of home computer companies. This restriction also made chassis design pretty simple for Felsenstein. These walnut boards were about 9 inches tall. That was one dimension done. The Sol single board computer was a long rectangular board, so that dictated the length of the computer. It seems to me that the final consideration, those being the overall profile and width, were pulled from Felsenstein's previous work. The Tom Swift Lives article,

the one that I talked about earlier, gives us the first glimpse of a wedge. The final case for the Sol 20 is almost identical to that diagram. Felsenstein drew heavily from this article, it would appear. At the front was an integrated keyboard that would dictate the width of the machine. The keyboard was mounted at a slight angle to make typing a little more comfortable. The walnut sides of the case were cut to roughly match this wedge shape, extending above the keyboard but keeping the slanted angle. Beyond the keyboard, the angle flattened out, forming a large rectangular back where the power supply and expansion bus were housed.

This was all enclosed in a bent sheet of steel, which was painted a distinctive blue. So here we have a working computer packed into a roughly wedge-shaped case. It has an integrated keyboard, a simple video output, and a microprocessor. We have arrived at a wedge. One interesting side product of this overall shape is the back portion, One interesting side product of this overall shape is the back portion, the block that houses the expansion cards and power supply. It can be used as a TV stand. So once a Sol is plopped down on your desk, you can stack a monitor right on top of the computer.

Proctech really did manage to build the most convenient computer possible at the time. You didn't really need anything extra, just a saw and a TV, which you probably already owned, and you were ready to blast into the computer future. I can't find the exact date, but a prototype was most likely ready in the first few days of January 1976. Felsenstein recalls that from idea to PCB was about 45 days, with some unspecified amount of time to get that from board to working machine. With the first mountain crossed, it was time to get back to Solomon at Popular Electronics. After a few last-minute modifications, the Solve20 prototype was ready to roll. Marsh and Felsenstein packed up the machine, caught a cheap red-eye flight, and made their way to the magazine's headquarters in New York.

As Felsenstein put it, This first meeting between Solomon and his namesake didn't really go very well. There was a short somewhere on the motherboard, so when it came time to actually demonstrate the machine, it just didn't work. Something was wrong with the display circuit, so the supposedly crisp text output was illegible. Felsenstein was eventually able to trace the error back to a loose piece of wire braid that was caught under a chip. The next day, he returned to Popular Electronics with a working computer, and the deal was finalized. Another little hiccup in these meetings was the whole terminal computer thing. Like I brought up earlier, Proctec was trying to keep away from the Altair's territory.

Solomon had asked them for an intelligent terminal, not a computer. But, I mean, come on. The Sol-20 is 100% a computer. Felsenstein makes it sound like he and Marsh were trying to keep that fact on the down-low and planning to pull back the curtain once the popular electronics cover story came out. Felsenstein even describes a point in the meeting where Solomon wasn't entirely convinced. Quote, Why couldn't he, he wanted to know, plug in a ROM board with BASIC burned in and run it standalone? I smiled my blandest smile and muttered, beats me,

end quote. They really weren't fooling anyone. Officially, it was billed as a really fancy terminal for the Altair or any other computer. But the Sol 20 was its own machine. And strictly speaking, you could plug in a ROM board and run BASIC on the machine. In fact, there was built-in support for ROM cards. The Sol 20 called them personality modules, but they're basically little circuit boards with ROM chips that could be plugged into a dedicated socket on the motherboard. This socket was accessible through an open slot in the case, so it would be, in fact, very easy to swap out the Sol's terminal software for BASIC. This brings us, quite handily, up to the computer's first big outing. In July 1976, the Sol 20 appeared on the cover of Popular Electronics.

It took eight months, just a little bit over the 30-day challenge we started with, but the machine had finally arrived. Sandwiched between ads for CB radios, turntables, and a new revision of the Altair, readers got their first view of a wedge-shaped computer. Titled BuildSol, an intelligent computer terminal. It keeps up the very thin facade of an intelligent terminal. But the article itself is somewhat underwhelming. All things considered, it's tiny. It's just a few pages long. It includes a parts list, in-depth specification for the system,

and some details on how to build the computer. Also, it has the all-important information on where to order circuit boards, schematics, and kits to build your own. At this point, you could order a full kit to build a Sol for $497. That's just a little more than $2,000 when adjusted for inflation. But keep in mind, this was a good deal cheaper than a decked out Altair. So I guess in that sense, we can start to see Moore's Law at play. In August, the Sol 20 made its next appearance, this time as a fully assembled computer, at the PC-76 computer fair in Atlantic City. Anyone who showed up expecting to see the machine

from Popular Electronics would have been a bit confused. The Sol 20 went through a major revamp between prototype and release. Nothing really changed that matters. It was still the same computer. The case was a little bit different, but it's mainly just touches and tweaks. The cover story had used photos of a really early and weird case. The motherboard circuits were retraced, connectors were moved around, that sort of thing. This fair was a pivotal moment for home computing in general. Felsenstein says it was the first personal computer fair, which may very well be true, I haven't found any earlier examples. Around 20 or

30 other vendors were there, most of the same stripe as Marsh and Felsenstein. Folklore holds that a group of Homebrew Computer Club members ended up on the same flight out to New Jersey, all with their handmade computers in tow. This is also where we run into a bit of contention in the story. So far, I've been claiming that the Sol 20 was the originator of the wedge computer design, or that it was at least the first wedge machine on the market. For me, that involved more than just the shape. It's also down to function.

Look at the early era of home computers, and then look at the Sol 20. You get an almost one-to-one match for every feature. The only notable difference is that the Sol 20 didn't originally ship with BASIC. The Apple II is a good example of this formula. It was shaped roughly like a wedge. The case shape is really surprisingly similar to the Sol 20's chassis. The back portion of the motherboard had slots for expansion cards. It had an integrated keyboard, and it output to a TV directly. The technology was different, but the broad brushes are almost identical. But this is advent of computing here. No episode would be complete without some complicating factor that

ruins a nice, simple story. The two big complications here come down to chronology and proof of influence. That is, one, was the Sol 20 actually the first home computer to follow this full formula? And two, did the Sol 20 directly influence later computers? It turns out there is actually someone close to this story that would say no to both of those questions. That someone is Steve Wozniak. Now, Apple history is always a little tricky for me to deal with. This is something that I find with any computer company that's still in active operation. Since there is still a functioning organization, many people connected will have some kind of agenda in how they tell a story.

That means you need to read with a bit of an extra-critical eye and be cognizant of your sources. The other issue with companies that are still running is you can't always find just an archive that's a huge dump of all their papers. For Apple history in this period, I usually reach for Steve Wozniak. He seems to be a much more reliable source than Steve Jobs, at least. Plus, Woz can offer more of the technical insights that I find interesting. And in this specific case, Woz is in a very interesting position. Apple, at least a very early version of that company, was in attendance at PC-76. They had a tiny booth where the Steves and a third friend

were showing off the new Apple I computer. In a 1992 interview with WGBH, Wozniak had this to say about the company's tiny machine. Quote, I knew the computer was the best. It was the one I would have wanted to buy. I mean, even the processor technology saw a computer, strangely enough. Lee might have felt to buy. I mean, even the processor technology Sol computer, strangely enough, Lee might have felt that then. But the Apple One was the first computer that came out that said,

a computer for this low cost for people should be a metaphor for a typewriter, kind of. It's got a TV screen, your home TV, and you type on a keyboard. That's how it comes. End quote. Now, it should be pretty clear given all the lead-up so far that Wozniak is talking about convivial cybernetics. He and Felsenstein are just using different language to express the same idea. Woz is also making the claim that the Apple One was hitting the wedge formula first. So, I guess we're going to have to start off by addressing the chronology issue. Waz was a member of the Homebrew Computing Club. He was involved in some of the first meetings all the way back in 1975. According to Iwaz, his autobiography, quote,

After my first meeting, I started designing what would later be known as the Apple I. It was inspiring. End quote. The Apple I was, all things considered, a capable computer. It was based off the Motorola 6502 processor. It generated its own video output signal, handled keyboard inputs, and even had an expansion interface for adding more RAM. However, this machine wasn't as plug-and-play as Woz makes it sound. Initially, the Apple I was sold as an unpopulated circuit board. A small number of pre-built computers were sold starting in July of 1976,

but these were still just circuit boards. A user had to provide a case, power supply, and keyboard. Pre-assembled Sol 20s started shipping in December of 1976. So yeah, technically, you could buy an Apple One before a Solve 20. But here's why I don't think that Apple wins this one. The lazy argument is that the Apple One wasn't really a complete package, not yet. You didn't have a computer that a consumer could buy, drop on a desk, and go. You didn't have a wedge that you could buy. But I think that's more of a nitpick than a really good argument. The guts are there, so fair enough. We can pull out a more compelling argument by thinking back to episode one of this series. Around the time the Homebrew Computer Club was formed,

Felsenstein was working on his Tom Swift terminal design. The machine was never built, but there's a paper trail. Crucially, Felsenstein was sharing his work at club meetings. He even specifically mentions talking about Tom Swift at an early meeting where Woz talked about his experience designing arcade boards. Tom Swift may have been a theoretical device, but it hit all the same beats that the Sol 20 and, for that matter, the Apple I were going to hit later. We can bulk up this argument some more. Going back to the article Tom Swift Lives, that has my key wedge diagram. That was written, or I guess drawn, in 1975. Once again, we don't have a month, but the year is good enough. We have a wedge diagram with an explanation of the right innards

pinned by Felsenstein in 1975. That design would go on to be mashed up into the Sol 20. So sure, Waz started working on the Apple I prior to the development of the Sol 20. But Felsenstein was on that same exact tip for years. And we have the paper trail to show it. The second question, the one of influence, is where I think things get a lot more interesting than just the raw chronology of this. The Sol20 was the first total package. I think we can lay that one to rest. But should we count the Sol-20 as the progenitor of the wedge computer? Would desks in this era look the same without ProcTech's little terminal? Just to be 100% fair, I might be looking for controversy where none exists. However, I think this gives us a

good framing for a larger conversation. Once again, we're sticking with the Apple comparison. The next computer to come out that really fulfills the whole wedge formula is the aforementioned Apple II. Looking at the two systems in profile, it's plain to see. The Apple II really does just look like a more refined Sol 20, at least on the outside. But correlation doesn't equal causation, so we have to look at this with a little bit more scrutiny. In the same WGBH interview, Woz admits that he didn't spend much time at PC-76 itself. For most of the fair, he was hunkered in a hotel room working on a prototype Apple II. He specifically says he was programming, so by the time he sees a Sol 20 in the flesh, the Apple II was already pretty close to completed. I think that's enough to say that

Woz wasn't just copying the Sol 20 to create the next Apple computer. But here's the weird part. Jobs may have been on a whole other wavelength completely, at least when it came to the look of the Apple II. At PC-76 and leading up to it, the demo Apple I was just crammed into a handmade wooden case. That was the norm for hobby computers at the time. Folk either used wood or sheet steel because it was something that you could deal with in a garage.

Jobs did not like either of those options. Proctech was only slightly higher up on the food chain. The Sol 20's case wasn't built out of someone's garage, it was painted, and it even had an official-looking name plaque. This may have been the only period where Apple was behind the curve on the whole visual polish thing. Jobs didn't like the Salt20's specific design, per se, but he did want a better case for the upcoming Apple II. As folklore goes, Jobs went out to a department store to gawk at appliances, trying to imagine how a computer could be made to look more at home in, well, the home. The Sol 20 looked too institutional for that, more like lab equipment than anything, and for some reason that is never articulated very well, Jobs just wanted to use molded plastic.

I'm not saying that just to be funny. He was quoted as saying, quote, I got a bug up my rear that I wanted the computer in a plastic case, end quote. Dude just wanted a plastic case. What more can we say? Jobs eventually contracted Jerry Manok, a fellow Homebrew Computer Club member, to do the final case design. Thus, we arrive at 1977 with the iconic Apple II, a plastic wedge with the space to stack a TV on top. Its exterior was designed all around the Sol 20. Everyone involved had contact with the Sol 20 in some capacity. Everyone involved had contact with the Sol 20 in some capacity.

So the plastic wedge is at least a response to the earlier sheet metal computer. So the outside of the Apple II bears a resemblance to the Sol 20. The internals have different details, but it hits all the same major features. All that said, I'm not suggesting that Apple stole some secret formula from Felzenstein. The Steves weren't quietly scheming and wringing their hands in the back of a homebrew computer club meeting. So, are we just looking at an oddly connected coincidence? I don't really think so. That doesn't explain all the other wedge-shaped computers that have expansion slots, cartridges, TV outputs, and all the other features that are in these two systems. For me, the best explanation is the most

simple one. I think this all comes down to trends and forces. In other words, I think the wedge was the next inevitable step. To wrap our heads around all of this, we need to go back a little ways and try to understand why Felsenstein arrived at the formula he did. We've explained the parts pretty well, at least I hope, but let's just run over them really quick so everything is fresh in our minds. Keyboard input is the easiest one. That was already well established as the most reasonable input method for a computer. It's a holdover from terminals. The TV output part is also pretty simple. Felsenstein wanted a cheap way to build a CRT-based terminal. TVs were cheap and available, and they could be easily replaced

when they broke. The Tom Swift terminal design brought these components together around a bus, which opened up options for expansion and eventually a microprocessor. Looking at just the tight, specific case, Felsenstein was trying to make a cheap and reliable terminal for community memory to use. But more broadly speaking, and simplifying things down, Felsenstein was trying to build a fun computer. As he put it, a convivial cybernetic device. For Felsenstein, a convivial computer was one that was easy to understand, easy to use, and easy to access. A device that could be owned and operated by a single casual computer enthusiast, not just nerds with degrees and soldering irons. As far as I'm concerned, that's a really modern way to look at computing. That's how we use computers today. When was the

last time you took a soldering iron to your desktop? In essence, Felsenstein was going for a totally different type of computing experience. That required a full retooling of what a computer was. The most obvious component, at least the most visible one, was the TV output. This gets into something that I glossed over a little bit when addressing the early days of the Altair. To use something like BASIC on the Altair, you had to use a terminal connected over a serial interface. Last episode, I harped on how that was a pretty costly setup, but it also just sucked. Hard copy teletypes, the kind that print outputs onto physical paper, are really limited machines. You can only get outputs line by line, which severely limits what you can do. Glass teletypes were also available, but even with a fancy CRT,

you still get throttled by the serial line. Even if you're on a local connection, even if you're only feet away from your Altair, everything had to come over at a set baud rate. That's the whole serial part of the serial connection. Data transfers only one byte at a time on a fixed and very slow clock cycle. You can't just throw data over to a terminal and call it good. You have to wait. The net result is limited throughput, both from the user to the computer and from the computer to the user. It means you don't get real interactivity. It also has a weird limiting factor on how quick your software can run. Felsenstein's VDM-1, his fancy video display module, helped shift away from that. It was transferring data from a chunk of memory straight to a TV.

All data stayed in memory, so the rate of transfer was high enough to seem instantaneous to us slow humans. So the rate of transfer was high enough to seem instantaneous to us slow humans. In an interview with the Computer History Museum, Felsenstein described his new video output method like this. Quote, The others were interactive, just not very real-time. So this opened up aspects of the personal computer. First of all, the play aspects, which I think are really critical, because most of what I do, at least, is a kind of play. Even if I'm being paid to do it, even if it's not something I was originally interested in doing, if I'm going to do it well, I have to be at play. And this ensured

that the personal computer was different from real computers. Real computers were used for business. They were used for making that interested in making serious computers. That ground was well tread already. The VDM-1 and, in general, the shift towards using TVs instead of terminals allowed for more interactive computers, and Felsenstein wasn't the only one going down this path. Just prior to the VDM-1, Don Lancaster's TV typewriter took the hobbyist world by storm. That device was so simple it was almost useless. All it did was allow a user to type characters on a keyboard and have them show up on a TV.

Lancaster even admits it was popular not because it was especially useful, but because it was fun. The shift to TVs and integrated keyboards also had a less technical impact. For one, it was just cheaper than any possible alternative. That was a huge goal for Felsenstein, and I don't think there was any other way to achieve his master plan. Real-time interactive displays did exist for computers, but those were highly specialized devices. You could even get a graphics display to connect up to your mainframe if you had the money. If you just needed a cheap way to display a little text, a TV was more than enough.

The circuitry to drive a TV signal was also pretty cheap, so you could save some more cash there. Crucially, this was recasting a computer as a combination of familiar devices. I haven't seen Felsenstein talk about this particular part explicitly, but this was an aspect that Jobs and Wozniak were all about. Woz called it the typewriter metaphor, basically the idea that a computer would come across to a consumer as a souped-up typewriter. Electric typewriters were already in the marketplace. People were comfortable enough with that product that it would be familiar to them.

Even if it did something new, it's just a fancier typewriter, right? Jobs had his own take on this. He talked about the Apple II specifically as a, quote, appliance computer. It was something that fit into the home, a machine that could be used just as casually as you use your refrigerator or washing machine. When planning the Apple II's case, Jobs specifically went looking at appliances to glean their overall look and feel. He wanted the new computer to fit into a home. You could put it on a desktop you already owned, plug it into a TV you already owned, and type on it just like a typewriter, but it could do a lot more than just print out characters on a page. Then comes the expandability part of the equation. This, I think, also has an interesting subtlety to it. The bus design that Felsenstein sketched out for the Tom Swift terminal solved a lot of

problems at once. It made his design easily expandable, but it also worked to lower the barrier to entry for a new user. We saw this with the Altair and the Sol20 and even the Apple II. A stock Altair can't do a whole lot, but it's so cheap. You get into the world of computing for a low upfront cost. You can buy into new features as the need grows. The Solve20 hit a really good stride where the base model had all the most common features you could ever want. Expansion modules let you add memory or a later floppy disk interface, maybe some more I.O. options. But you could get up and running without wasting money on features that you didn't want or maybe didn't even know existed. The Apple II followed a really similar route. A base model had video output, a cassette tape interface,

keyboard input, a processor, and some RAM. If you wanted floppy drive support, then you could buy into that feature somewhere further down the road. Same for more RAM or even a different processor. While that all cost more money, it meant that a base model could be a lot cheaper. A computer was still a huge investment, but it was more on the scale of a new refrigerator instead of a new car. Alright, that brings us to the end of this episode, and I guess that also

brings us to the end of a month-long dive into the Sol 20. But more broadly, this was also a look at how home computers in the 70s and 80s took on a distinctive flavor. In classic fashion, the big point I want to leave you with is that the story of home computing is complicated. In a roundabout way, I think the final form that we saw on desks in this era was almost inevitable. In a roundabout way, I think the final form that we saw on desks in this era was almost inevitable. Felsenstein wasn't the only one hurtling towards a wedge-shaped computer with an integrated keyboard, video output, and simple expansion options. There was a whole cohort of hobbyists around the same era thinking along the same lines.

Each choice along the path to the design of the Sol 20 was heavily influenced by the era Felsenstein was working in. In the end, it wasn't massive corporations that broke into the home market. It wasn't a single genius with a new idea. It took time, it took passion, and it took a community. Ultimately, it took nerds in a garage that wanted to share their brand of fun with the world. Before I sign off, I just want to make a quick announcement. Right now, I'm running a poll over on Patreon for a new bonus episode. It's going to be coming out to patrons sometime in September, so if you want to get in on that and support the show, then why not head over to Patreon and sign up? Right now, I have, I think, four bonus episodes in the catalog so far, and this should be the fifth. Anyway, I'll be back in two weeks with another full-length episode covering the story of computing's past. Links to everything

are up on my website, adrenofcomputing.com, and if you have any comments or suggestions for a future episode, then go ahead and shoot me a tweet. I'm at adrenofcomp on Twitter. Thanks for listening, and as always, have a great rest of your day.