• Please review our updated Terms and Rules here

Thoughts on minicomputer to microcomputer transition (PDP to Macintosh)

One other critical influence on the development of the early microcomputer was Don Lancaster's TV Typewriter with its resolution of 32x16. Complete the entire book and the guidelines would provide enough information to have working keyboard, display, and cassette interface. Just add a CPU and memory, make some tweaks to cut costs and improve resolution, and very quickly one had the basis for an affordable microcomputer.
 
Eudimorphodon said:
That’s a claim I’ve never heard before. Citation needed?
Click to expand...
Regarding 80 vs 40 col PET... In the 1977 release, it was all 40-col (i.e. their original 2001-series models with the smaller keyboard). But in 1980, the CBM 8032 and PET 4032 were apparently released at about the same time in May 1980.
http://oldcomputers.net/pet4032.html
Somewhere in the Facebook group about the PET, some original owners noted the 80-col was technically sold a week or month before the 40-col. I couldn't find any Commodore brochures or catalogs to verify that, maybe it was based on their sales receipts.

That Trinity collectively represents the "PC-as-an-countertop-appliance" era (to me). The software wasn't quite there in 1977, but by 1980 these had become viable systems for a few home applications and small business. Old-time General Dynamic employees told me stories how in 1978, they used the PET during the design of the F-16 OFP software (because they didn't need the color capabilities of the more expensive Apple, and the PET had floating-point whereas Apple's initial INTEGER BASIC did not).
 
Last edited:
voidstar78 said:
Somewhere in the Facebook group about the PET, some original owners noted the 80-col was technically sold a week or month before the 40-col.
Click to expand...

But that's three years after the PET came out, so the 80 column PETs were certainly not out first.
voidstar78 said:
Click to expand...

So, a comment about the info on this page: long and short of it, it's not really complete/accurate. The original small-keyboard PETs were joined by the "2001-N" in 1979 (not 1980), which included the new keyboard, a new motherboard that used dynamic RAM instead of the SRAM of the original boards, and fixed the video display RAM access so it could be done on the off-cycles of the 6502, thereby "speeding them up". (The original PETs would get snow on the screen if VRAM was accessed outside the VBLANK area, so BASIC tracked a refresh interrupt for screen updates.) This "2001-N" model was sold under a bewildering welter of names depending on what region you were in (or whatever stickers somebody ordered that day), so you'll find them with "2001-N", "3008-3032", "4008-4032", whatever, on the label. The number *sometimes* has some correlation to what BASIC version is installed, but that's about it, they made the same machine for quite a while.

What came out in 1980 was the first 80 column PETs; these had a new motherboard that differed from the old board in that it had a CRTC chip on it (replacing a pile of discrete video timing logic) and they drove a 12" monitor instead of the 9" one. For a brief period this CRTC board was only used for the 80 column PETs, but within a few months they modified it into a "Universal" board which could be used for either 40 or 80 column displays (when configured with different ROMs and quantities of VRAM). This is when 12" 40 column PETs showed up.

So... yeah, I guess, sure, I guess if you want to be really specific and say the 80 column 12" Monitor PETs came out before the 12" 40 column ones then that's true, but it's also pretty meaningless. There's almost no functional difference between a 40 column 12" PET and a nine inch one if they're running the same BASIC version. Which they can do; "4032s" came from the factory in both sizes and they run the same software at identical speeds. The only difference is the 12" models are prone to be melted by "Killer Pokes", the 9" ones, lacking CRTCs, are immune.
 
Right, I misspoke, I meant "first" in the subsequent updates after the 2001-series (when they went to the full size keyboard models). But the note about the "universal" board is interesting, in that it wasn't a huge deal to adapt a 40-col model to an 80-col (or vice versa) (not a simple switch, but perhaps a single ROM exchange) - and explains why the 80-col did come out first for that series. They probably intended to make an 80-col model going to the larger screen, but then re-considered after realizing the ramifications that would have on existing PET software - so users had a choice.

EDIT: And yes, in magazine articles from 1983, you still see Commodore advertising those 4032 models. The SuperPET was just a bit late, and 96KB expansion wasn't going to cut it - the design of an attached screen was cute in 1977, but didn't stand the test of time. It was a good idea for schools, so students didn't walk off with a keyboard or a screen (and logistically it was easier for offices and schools to order a single part number, so a true all-in-one solution like the PET or TRS-80 Model 3 made that easier). But for the home environment, the weight/bulk of a full single PC wasn't practical. Commodore leaned towards the gaming market, then 1983 was the big console crash; Jack Tameriel departed in 1984 and pursued a deal with Atari (while Commodore redirected with the Amiga line).
 
Last edited:
The TV Typewriter is interesting, I hadn't come across those. I was familiar with the "Green Thumb" project that was a few years after that, it was some collaboration in Kentucky with the Weather Service wanting to commission a way for farmers to use a modem to get weather and farming-related information (with the VideoTex being a pre-cursor to the CoCo). It failed since the chip complexity and cost was high - but similar to the TV Typewriter project, it seems the main contribution of the effort was helping to evolve the text screen output (via the MC6847 integration).

Those kinds of contributions are certainly part of the story, and the various kits available motivated a lot of ingenuity of what could become possible with that new tech. But yes, fitting all those intermediate details needs a much larger image :D

In the original mural image, I did concentrate more on the "mass-produced" products. The KINBAK-1 was kind of an exception, I don't think it became commercial successful (I believe I had read something like only 50 were produced). I think the Micral did have some success locally/regionally in France.

Production numbers would be interesting to estimate. I don't think the Sol-20 was ever mass produced either - being sort of like the Apple-1 (i.e. not a production-line type of production). I think each of the HP9830, Wang 2200, IBM 5100 were easily in the thousands per year (but I've never come across exact production numbers). But each of the 1977 Trinity were easily in the 10,000 of thousand per year.

But on the other hand, those kinds of numbers don't apply to mainframes: a single mainframe could handle dozens (if not hundreds or maybe thousands) of terminal connections.

I certainly agree the PDP-1 marked a turning point towards making computers more personal. So the 1960s were a long transition from the million dollar mainframes to $100k systems. Then the 1970s was getting those $100k systems down to $1k (and now-a-days we have those $10 ARM sticks with BT/WiFi!). I guess one relatable analogy is like the transition from locomotive to an automobile, with decades of innovation needed (and similarly, today will still have "large transportation" vehicles along with very refined personal-transports).
 
Last edited:
voidstar78 said:
Anyhow, I'm still trying to understand the PDP-8 vs PDP-10/11/12 better.
Click to expand...

Originally DEC made different incompatible PDP models to address different market segments. At the beginning each new computer got a different PDP-# designator. With the 8 and 11 they stopped doing that since they were mostly compatible. PDP-8 was always the lowest cost lowest performance machine. Each new model was cheaper but performance didn't change much. It was popular in education, small business type stuff, and embedded application.. This or the PDP-12 would be the most personal of the machines you listed. The PDP-11 was the next step up and did significantly increase in performance with new models. PDP-11's were more likely to be multiuser machines and much of Unix development was done on it. 8's did have multi user option. PDP-10 is big multiuser system. PDP-12 was the end of the line of the dual instruction set machines which implemented the LINC instruction set and the PDP-8 instruction set. LINC was LINC only, LINC-8 was LINC and straight 8 processor, PDP-12 was LINC and PDP-8/I processor. 8 and 12 are 12 bit, 11 is 16 bit. 10 is 36 bit.
 
Intertec, not to be confused with Intellec ?

I wasn't familiar with the Intel's Intellec (1973) until I read thru Gary Kildall (CP/M) notes "Computer Connection" Outside, the Intellec looks basically the same as an Altair 8800 (but push buttons instead of flip switches?) -- and it has some sort of bus inside, but not S-100 (at least I don't think so?). Not sure why Intel's kit was ~10x the cost of the Altair.

Gary mentions the MOS 6502 as being "horrible" and suggests that the Softcard for Apple-2 "saved" both Apple and Microsoft c.1979 :D Maybe.
 
I never used a PDP, but I was curious to try to summarize the variations. For each main model, there are lots of "trims" or sub-models, with variations configurations. So it's hard to sometimes have a single value for each main model (and I gave up trying to summarize processing "speed" and memory-capacity). Summarizing the info from Wiki, the general story about the PDP's looks like this:

1663799191708.png

Some articles seem to peg the PDP-5 as the first practical personal computer in 1963 (I think since it was a single cabinet; the medium home cost in 1963 was $18k - so conceivably someone possibly could have financed a $27k computer back then -- and it's the first to reach 1000+ units sold). [ Gary Kildall also explains why that debugger was called DDT {as in Raid} ]

I put the PDP-4 as the "alpha" version of Unix and B. I don't know the full story there - initial development maybe started there (hence "alpha"), but other articles say the first "official" Unix wasn't until 1969/1970 on the PDP-11.

I don't have a "weight" for the PDP-11, since there just so many configurations.

And I don't have a "cost" for the PDP-10, I suspect it was back over $100k?

All this, of course, is just a piece of the mainframe to minicomputer story. Certainly more "personal" than the mainframes, but a minimum of 200+lbs and still no standard "character generator" screen (to do away with those screaming line printers), not yet the home appliance type device.
 
PDP-11 has a lot of models. The model numbers indicated the maximum memory, which bus was used, and whether TTL or IC was used for logic. Only a few models were sold at any time so it was easy to do a comparison of the offers.

Important models would be the original /20, /45 (256 K), /70 (4 MB), and low cost IC equipped /03 (64 K) which was followed by going to higher speed ICs /23 and /73 restoring the 256K and 4MB RAM limits.

Gordon Bell's website has some interesting evaluations of the PDP-11. http://gordonbell.azurewebsites.net/Digital/DECMuseum.htm Also, see the interview towards the end with Ken Olsen which covers the single user aspects of MIT's TX-0 (which led to the PDP-1).
 
Updated "mural" (just for casual for fun musings):
1664095719354.png

No single "fits on a page" diagram can show all the nitty-gritty details here, still focusing on select highlights.

- Added the PDP-8/E. PDP was great, but still over $6000 base, so still a need to reduce cost and less use of a line printer as output.
- Added Magnavox Odyssey. It's a bit out of place, but I think it did help contribute towards that of idea of just using a TV as your screen (and more compact hardware to help make that happen). It derives from the "Brown Box", but no room for the subsequent 1975 home-version of Pong (and didn't want to include the 1972 arcade tower of Pong).
- Added Intellec, even though it wasn't really a consumer product but explicitly a dev kit from Intel. (the Micral N also used an 8008)

- Added Intertec, Model 3, OSBORNE, and ColecoVision along a clearer "Z80" lineage line. I couldn't find an original user guide to the Intertec (did it run CP/M?). ColecoVision is also a little out of place, since it wasn't a re-programmable device, but it shows how a decade later (from the Odyssey) game systems were still right in the mix of the micro-development.

- Added 400/CoCo/Micro along a "6502" line (CoCo used a 6809).

- Added more software title context. I realize Electric Pencil predates WordStar, but I couldn't find a "screenshot" of a pre-1980 Electric Pencil (4K systems had no spare memory to be showing load titles!). But I wanted to show that while the word-processing and cell-calculator were the initial killer apps (WordStar, VisiCalc) -- but micros quickly brought in a lot of other applications that evolved greatly between 1978 and 1980.
-- Microchess early version ran on the original KIM-1.
-- No tape version of Zork exists, english-text quickly consumes space, so it needed a large-capacity disk to be released.

- Added Micromodem and PETUNIA as examples of expanding the capability of the base system. (#1) The Micromodem 300bps was first available for the Altair (S-100 bus), quickly followed by supporting the Apple bus system. Someone earlier mentioned communication between systems being important - it took all the resources of these early systems just to run a terminal program and asynchronously maintain the connection, so online gaming wasn't quite happening yet for micros (especially not with just 4K RAM). Still, the capability was in its infancy [ plus AT&T had a stranglehold, in short making use of modems difficult until ~1978 ]. (#2) PETUNIA will be very little known, but it represents a first "sound card" of sorts (and as an example of software that was delivered on tape).


I also wanted to add a screenshot for KERMIT, as one of the earliest terminal emulation programs for the IBM PC (1981), but couldn't really find one (I did find the old user manual). Again, startup titles of software cost resources, so a lot of early software just didn't have one.

And for the IBM PC, I wanted to add a screenshot of the early title called SOPWITH - which demonstrated two player gaming (via an interlink cable) on the IBM PC (and while using CGA graphics). Does anyone recall if the Apple II (with only 48K or less) had a similar demonstration program? (i.e. one machine talking to another, so two players coordinated across a connection)
 
There was a game called COMMBAT which ran on Apple, Atari, and TRS-80.

www.mobygames.com

Commbat (1981) - MobyGames

Commbat is a two-player combat strategy game that is played via two separate systems connected by modem.War has been declared between two nations regarding the mining rights of the third planet of the Deneb star system. In accordance with Galactic Law, this issue shall be settled in single...
www.mobygames.com www.mobygames.com
 
Bruce Tomlin said:
There was a game called COMMBAT which ran on Apple, Atari, and TRS-80.
Click to expand...

Neat, hadn't seen that one - it definitely showcases the potential. On the box it says 24K and needs a DOS (i.e. disk release). Whatever DOS it is referring to on the Atari system, that takes up some resources - so the applications leverage those resources to do things (so the size of the application appears smaller, but it is probably using 4-8K of "system calls" that the DOS is providing -- like disk IO routines). Also, I see the in-game menu list a (S)AVE feature (in contrast, a few tape-titles had a save feature -- I can only think of barely a handful).



So, "personal computer" is a long gradient between 1960 to about 1984. It depends on where you put the thresholds on cost, weight, and utility (such as whether vector graphics or a DOS are essential standard part of the system). Those factors influenced the "popularity" of the system (to attract a bulk of developers).

(e.g. I've considered the Player Piano from the turn of the (last) century as a form of the first personal computer -- with 88 "bits" {keys} of "vibration memory" and punched paper as the software; they were affordable, but still heavy and just one application of emitting tones - but no RTC and limited floating point precision :p )
 
Last edited:
I'm not sure that this would be classed as a mini-computer at the time, and I don't remember which company made it (the only thing I remember is that the computer was a 12bit one). And this is just third hand knowledge from a co-worker that I worked with in the late 80's. He worked with a computer that you didn't outright purchase, but instead leased from the manufacturer. The computer case was locked and only the representatives from the company had a key. When they wanted to upgrade the computer to a faster version, the rep came by, unlocked the case, removed long cables, and then put in short cables. That was the extent of the upgrade. I don't think that would be considered a "personal computer" in this context. :)
 
Shorter cables :D Well, I recall someone made a "monster" (enlarged) version of a MOS 6502 CPU. Two things stood out to me about that: #1) it was functionally equivalent to a 6502, but it ran substantially slower just by virtue of having longer wires to communicate the signals and #2) would be a neat to have a museum with a "room sized" CPU that you could walk thru, and see Instruction Fetch and execution in action. Perhaps VR will give us something like that in the future. [ or perhaps mainframes way-back were essentially like that, you literally walked inside the computer ]
 
Last edited:
grimm said:
I'm not sure that this would be classed as a mini-computer at the time, and I don't remember which company made it (the only thing I remember is that the computer was a 12bit one). And this is just third hand knowledge from a co-worker that I worked with in the late 80's. He worked with a computer that you didn't outright purchase, but instead leased from the manufacturer. The computer case was locked and only the representatives from the company had a key. When they wanted to upgrade the computer to a faster version, the rep came by, unlocked the case, removed long cables, and then put in short cables. That was the extent of the upgrade. I don't think that would be considered a "personal computer" in this context. :)
Click to expand...
There was a model of the CDC CYBER mainframe that had an extra wait state inserted to make it slower, allowing for a lower purchase price. Some enterprising CEs offered to remove the jumper in exchange for gratuities until management got wind of the practice.
 
We had a similar issue with the bean counters with the Durango 900 series. Initially, it used Rodime 7 and 14MB hard drives and was priced accordingly. Rodime discontinued those models after a few months, shipping 10 MB and 20 MB units instead. We had to add code to limit the 10 and 20s to 7 and 14 MB; marketing added another couple of lines in the pricing guide for a 10 MB and 20 MB "upgrade", which entailed a different boot disk. I'm sure that similar stories can be told.
 
You must log in or register to reply here.
Back
Top