(Note: Yes, I know a 3B2 is not related at all to a 3B1! But I figure if you're interested in one, you're probably interested in the other.)
As late as 2002, at AT&T Easylink Services we were still using 3B2s to process email and email-to-fax on a private X.25 network; the beginnings of a 3B2 emulation-on-Sun-hardware project was in the works; I don't know if it ever went anywhere.
I originally thought this was an emulator to run (original) AT&T Unix applications on a PC. Then I googled it and found the article you linked. The photo in the article doesn't really give you a good idea of how the system looked like when in use, here's a better one: https://oldcomputers.net/pics/att-unix-pc-left.jpg . If the keyboard is attached to the main case, the design makes it look so flush that you can overlook the fact that it's detachable.
Like, MS-DOS? How? It doesn't seem to have had an x86 CPU to run it on. (I'm assuming this predates emulation like we might use today for that kind of thing.)
Edit: Oh, kept reading the wikipedia article and found the mention of an 8086 expansion card. That's cool:)
We had PC emulators on the 68000 based Atari ST in the 80s. No extra hardware needed, but very slow! And I'm sure that emulators was a thing from the very first computers.
Some systems also had expansion cards with an x86 processor on them. A more recent example would be Sun's SunPC and SunPCI cards -- you plugged these in a SBus or PCI slot, and ran x86 code on a real x86 processor. IIRC that's actually what the AT&T Unix PC did, too, but I'm not sure (I never owned one).
I had one of these in my Amiga 500: https://www.edsa.uk/blog/the-kcs-power-pc-board
(this was before the PowerPC CPUs appeared on the market, so the name wasn't as confusing as it sounds today).
In addition to being a PC board, it also doubled as a memory expansion and RTC when you were using the host Amiga system.
The ST was at least a cost-effective way to emulate a PC very slowly. Not so much for SoftPC for VMS, which cost as much as an entry-level ST for the software alone:
But hey, MS-DOS 6.2 and Windows 3.1 would have at least both fit comfortably on a single TK50 cartridge or 9-track open-reel tape, so at least you weren't stuck in front of the console swapping floppies waiting for the products to install.
That must've been it! My friend and I bought them used at a yard sale at the end of the airport from a couple who had worked at TWA - they were former TWQ machines. I don't think we ever opened them.
I have a 7300 and a 3B1 (7300 is single half-height hard drives, 3B1 is two HH or one full-height). One has a Combo Card (RAM and serial) and a Floppy Tape card, I can't remember what the other has, but cards have been pretty thin on the ground. A DOS-73 or network card would be extremely nice to have.
The emulator actually started life because I couldn't find a machine for sale at a price I could afford, but the manuals were on Bitsavers. Someone later sold me the 7300 and 3B1.
You could connect a phone to it. I think I saw an old magazine article (maybe BYTE) about it once, which said that, and had a photo of it, with the phone.
The Wikipedia article article says it had 3 phone jacks.
Yeah, back in the 80s, "PC" wasn't synonymous with "IBM PC/clone" yet. That only happened by the mid/end 90s, when all "personal computer" competitors (except for the Mac) had been forced out of the market...
Being Australian, I don't know much about the company other then it's a popular telecom provider in North America so it's interesting to see that they had their hands in workstations at some point.
AT&T was the renamed Bell Telephone, responsible for the invention of huge swaths of technology. You can look at old PDFs of the Bell System Technical Journal, published 1922 to 1983. Bell was also the birthplace of Unix.
”Researchers from there are credited with the development of radio astronomy, the transistor, the laser, the photovoltaic cell, the charge-coupled device (CCD), information theory, the Unix operating system, and the programming languages B, C, C++, S, SNOBOL, AWK, AMPL, and others. Ten Nobel Prizes and five Turing Awards have been awarded for work completed at Bell Laboratories.”
AT&T was the telephone monopoly in the US until the government broke it up into 9 regional companies in the 70s. These companies have since re-merged into 3 companies: AT&T, Verizon and Lumen Technologies. Lumen is the only one that isn't in the cell phone business. It used to be called CenturyLink until very recently.
In fact, AT&T agreed not to enter the computer market in return for being allowed to maintain a monopoly. They agreed to the 1984 breakup only in return for being allowed to make computers - a decision which did not go well, since their computer line was never very successful.
In terms of strategy it's not a bad move though, after all the computer (and by extension the smartphone) did become the new telephone. Pretty much all phones even run an OS with Unix-like underpinnings.
They just failed to capitalise on it but it was a pretty good vision IMO.
In 1995 I was an intern at Indosat, a telephone company in Indonesia.
Indosat bought billing system from ATT and also some PCs from them (maybe the PCs were part of the deal).
That is how I experienced the speed of a Pentium PC for the first time.
The machine itself was made by Convergent Technologies. It's similar to the MiniFrame - and possibly also the MightyFrame.
AT&T rebranded it, and there were two versions: the 7300 can take a single half-height MFM drive, and the 3B1 can take two half-height MFM drives or a single full-height. The 3B1 has a different plastic case with a square-ish 'bump' under the monitor, and usually a modified (called a P5.1) motherboard.
They top out at 4MB RAM (2MB on the motherboard and 2MB on expansion cards). Disk storage would have topped out at 190MB with a single Maxtor XT-2190 full-height, or a pair of Miniscribe 3650s for 2x50 = 100MB.
While I have heard about the 3B1 for years, even if AT&T is an unknown foreign company in my part of the world, I'd never heard of the MiniFrame. Thanks for this.
Like pretty much everything Convergent did, it was ahead of its time. Some really brilliant people must have worked there. Too bad so few people has hear about them (or used their stuff).
The telltale sign of late Convergent is the keycaps, also present in many Burroughs and Unisys desktops and terminals.
The 3B1 was one of the few uses of the 68010, which was basically a 68000 with a few bug fixes and, importantly, support for virtual memory. Combined with a separate MMU, this allowed you to run Unix.
The VM support was the killer feature. The 3B1 has a hardware (discrete logic!) MMU, a custom designed thing completely unlike the Motorola 68451. It's an odd beast and took a lot of work (by several people, not just myself; check the git commit logs and CREDITS) to get it right.
Another neat aspect of this machine was the hype around it at the time. AT&T took out ads in most magazines and pundits thought it would overtake the IBM PC because it ran UNIX and had some great software, including the first version of Microsoft Word (which was originally written for UNIX, specifically Microsoft's Xenix).
It's really a shame how little of that software has survived. I think I have a complete Foundation Set disk set, some others I wrote from ImageDisk images, and that's about it.
UNIX software in general seems pretty thin on the ground. The OSes are out there (even rarities like Interactive Unix 4.1 - though not the later patches e.g. FDISK 2GB) but software? Hen's teeth.
ACCELL Unify would be interesting to find, it's the software General Instrument used to write the front-end for their cable TV headends.
Ah, that's neat - it'd be nice to have it for SCO or Interactive Unix on x86. Especially Interactive, as that's the one I have to try and fix (Y2K20 issues).
Back in the early 90's, I had a friend with one of these. It was a neat machine. I remember an unusual windowing system on it ("MGR", I think?) We were starting to get into PC unix clones - first Coherent, then Linux (SLS distro?) - around the same time.
MGR was a later one - it could run (slowly) with an unmodified machine but really needed a VIDPAL to be used best. VIDPAL was a replacement PAL which enabled user-mode access to the framebuffer.
As an intern during college I remember installing the OS on a 3B1 at one of my first summer jobs. So many floppy diskettes. Don't remember much more about it other than it seemed pretty primitive compared to the Sun workstations that were prevalent around the office.
And probably the only real at-scale deployment of the Western Electric 32000 (Bellmac 32) processor. Sadly that chip wasn't very popular.
As LeoPanthera said above, Bell's exchange for the ability to make PCs was a bit of a bad swap in retrospect. It's pretty easy to see they were going for some kind of convergence of computing and telecoms (the 3B1 has an internal 1200 Baud modem) but they didn't do well on that bet.
(Note: Yes, I know a 3B2 is not related at all to a 3B1! But I figure if you're interested in one, you're probably interested in the other.)
As late as 2002, at AT&T Easylink Services we were still using 3B2s to process email and email-to-fax on a private X.25 network; the beginnings of a 3B2 emulation-on-Sun-hardware project was in the works; I don't know if it ever went anywhere.
Some interesting documentation I wrote here: https://3e.org/private/gms/
For those who didn't know what this was and (wrongly) thought it was some sort of IBM-compatible: https://en.wikipedia.org/wiki/AT%26T_UNIX_PC
I originally thought this was an emulator to run (original) AT&T Unix applications on a PC. Then I googled it and found the article you linked. The photo in the article doesn't really give you a good idea of how the system looked like when in use, here's a better one: https://oldcomputers.net/pics/att-unix-pc-left.jpg . If the keyboard is attached to the main case, the design makes it look so flush that you can overlook the fact that it's detachable.
> thought this was an emulator to run (original) AT&T Unix applications on a PC
Technically it is, if you mean System V for MC68k :)
But it sorta was! I had a 3B1. You could run DOS inside a window in the main environment.
Like, MS-DOS? How? It doesn't seem to have had an x86 CPU to run it on. (I'm assuming this predates emulation like we might use today for that kind of thing.)
Edit: Oh, kept reading the wikipedia article and found the mention of an 8086 expansion card. That's cool:)
We had PC emulators on the 68000 based Atari ST in the 80s. No extra hardware needed, but very slow! And I'm sure that emulators was a thing from the very first computers.
Some systems also had expansion cards with an x86 processor on them. A more recent example would be Sun's SunPC and SunPCI cards -- you plugged these in a SBus or PCI slot, and ran x86 code on a real x86 processor. IIRC that's actually what the AT&T Unix PC did, too, but I'm not sure (I never owned one).
Sun also briefly made a 386 based workstation. It didn't do well, but it could run DOS via V86 in a window on an otherwise Sun desktop.
https://en.wikipedia.org/wiki/Sun386i
See "DOS-73 Technical Reference" on the linked page for details. Yes, DOS-73 expansion cards had 8086 CPUs.
Amigas as well, that is how some of my friends would do their assignments on high school.
Yeah and one of the reasons was that every byte had to be reversed due to the endianness. So it was never going to be anything but slow.
There were pc addon boards for this reason for both Amiga and Atari but they were expensive as they were basically a whole pc on a board.
I had one of these in my Amiga 500: https://www.edsa.uk/blog/the-kcs-power-pc-board (this was before the PowerPC CPUs appeared on the market, so the name wasn't as confusing as it sounds today).
In addition to being a PC board, it also doubled as a memory expansion and RTC when you were using the host Amiga system.
Mine had an 8086 too!
The ST was at least a cost-effective way to emulate a PC very slowly. Not so much for SoftPC for VMS, which cost as much as an entry-level ST for the software alone:
http://bitsavers.trailing-edge.com/pdf/dec/catalog/Digital_U...
But hey, MS-DOS 6.2 and Windows 3.1 would have at least both fit comfortably on a single TK50 cartridge or 9-track open-reel tape, so at least you weren't stuck in front of the console swapping floppies waiting for the products to install.
Ah, you had a DOS-73 expansion card?
That must've been it! My friend and I bought them used at a yard sale at the end of the airport from a couple who had worked at TWA - they were former TWQ machines. I don't think we ever opened them.
Nice setup!
I have a 7300 and a 3B1 (7300 is single half-height hard drives, 3B1 is two HH or one full-height). One has a Combo Card (RAM and serial) and a Floppy Tape card, I can't remember what the other has, but cards have been pretty thin on the ground. A DOS-73 or network card would be extremely nice to have.
The emulator actually started life because I couldn't find a machine for sale at a price I could afford, but the manuals were on Bitsavers. Someone later sold me the 7300 and 3B1.
You could connect a phone to it. I think I saw an old magazine article (maybe BYTE) about it once, which said that, and had a photo of it, with the phone.
The Wikipedia article article says it had 3 phone jacks.
It's a perfectly good personal computer, though!
Yeah, back in the 80s, "PC" wasn't synonymous with "IBM PC/clone" yet. That only happened by the mid/end 90s, when all "personal computer" competitors (except for the Mac) had been forced out of the market...
Being Australian, I don't know much about the company other then it's a popular telecom provider in North America so it's interesting to see that they had their hands in workstations at some point.
AT&T invented, among other things, transistors, lasers, photovoltaic cells, CCDs, information theory, radio astronomy, C, and Unix.
C++ too, right? Unless Bell Labs was spun out of AT&T before that. Stroustrup started C++ at Bell Labs as C with Classes.
https://en.m.wikipedia.org/wiki/C%2B%2B
We are only mentioned the good things they did. ;-)
S and thus R came out of there too.
AT&T was the renamed Bell Telephone, responsible for the invention of huge swaths of technology. You can look at old PDFs of the Bell System Technical Journal, published 1922 to 1983. Bell was also the birthplace of Unix.
And then some:
”Researchers from there are credited with the development of radio astronomy, the transistor, the laser, the photovoltaic cell, the charge-coupled device (CCD), information theory, the Unix operating system, and the programming languages B, C, C++, S, SNOBOL, AWK, AMPL, and others. Ten Nobel Prizes and five Turing Awards have been awarded for work completed at Bell Laboratories.”
https://en.wikipedia.org/wiki/Bell_Labs
AT&T was the telephone monopoly in the US until the government broke it up into 9 regional companies in the 70s. These companies have since re-merged into 3 companies: AT&T, Verizon and Lumen Technologies. Lumen is the only one that isn't in the cell phone business. It used to be called CenturyLink until very recently.
Unix and C were invented at AT&T's Bell Labs.
https://en.wikipedia.org/wiki/Regional_Bell_Operating_Compan...
In fact, AT&T agreed not to enter the computer market in return for being allowed to maintain a monopoly. They agreed to the 1984 breakup only in return for being allowed to make computers - a decision which did not go well, since their computer line was never very successful.
In terms of strategy it's not a bad move though, after all the computer (and by extension the smartphone) did become the new telephone. Pretty much all phones even run an OS with Unix-like underpinnings.
They just failed to capitalise on it but it was a pretty good vision IMO.
Which was also when the lawsuit against BSD came to be, and the attempt to forbid the Lion's Books to keep being printed.
Which only goes to show what have happened to UNIX outside Bell Labs if it was a commercial product from day 1.
In 1995 I was an intern at Indosat, a telephone company in Indonesia. Indosat bought billing system from ATT and also some PCs from them (maybe the PCs were part of the deal). That is how I experienced the speed of a Pentium PC for the first time.
The machine itself was made by Convergent Technologies. It's similar to the MiniFrame - and possibly also the MightyFrame.
AT&T rebranded it, and there were two versions: the 7300 can take a single half-height MFM drive, and the 3B1 can take two half-height MFM drives or a single full-height. The 3B1 has a different plastic case with a square-ish 'bump' under the monitor, and usually a modified (called a P5.1) motherboard.
They top out at 4MB RAM (2MB on the motherboard and 2MB on expansion cards). Disk storage would have topped out at 190MB with a single Maxtor XT-2190 full-height, or a pair of Miniscribe 3650s for 2x50 = 100MB.
While I have heard about the 3B1 for years, even if AT&T is an unknown foreign company in my part of the world, I'd never heard of the MiniFrame. Thanks for this.
http://cholla.mmto.org/computers/miniframe/index.html
https://www.computinghistory.org.uk/sec/2963/Convergent-Tech...
The original brochure for the MiniFrame Plus:
https://archive.org/details/bitsavers_convergentmePlusProduc...
A blog about its big brother the MightyFrame, with pics:
http://mightyframe.blogspot.com/
Like pretty much everything Convergent did, it was ahead of its time. Some really brilliant people must have worked there. Too bad so few people has hear about them (or used their stuff).
The telltale sign of late Convergent is the keycaps, also present in many Burroughs and Unisys desktops and terminals.
The 3B1 was one of the few uses of the 68010, which was basically a 68000 with a few bug fixes and, importantly, support for virtual memory. Combined with a separate MMU, this allowed you to run Unix.
The VM support was the killer feature. The 3B1 has a hardware (discrete logic!) MMU, a custom designed thing completely unlike the Motorola 68451. It's an odd beast and took a lot of work (by several people, not just myself; check the git commit logs and CREDITS) to get it right.
Another neat aspect of this machine was the hype around it at the time. AT&T took out ads in most magazines and pundits thought it would overtake the IBM PC because it ran UNIX and had some great software, including the first version of Microsoft Word (which was originally written for UNIX, specifically Microsoft's Xenix).
It's really a shame how little of that software has survived. I think I have a complete Foundation Set disk set, some others I wrote from ImageDisk images, and that's about it.
UNIX software in general seems pretty thin on the ground. The OSes are out there (even rarities like Interactive Unix 4.1 - though not the later patches e.g. FDISK 2GB) but software? Hen's teeth.
ACCELL Unify would be interesting to find, it's the software General Instrument used to write the front-end for their cable TV headends.
> ACCELL Unify would be interesting to find
Both WinWorld[1] and Vetusware[2] have (what look like identical copies of) it for the 3B2.
[1] https://winworldpc.com/product/unify/31
[2] https://vetusware.com/download/Unify/?id=11459
Ah, that's neat - it'd be nice to have it for SCO or Interactive Unix on x86. Especially Interactive, as that's the one I have to try and fix (Y2K20 issues).
I owned a 3B1 back in the early 90's. It hosted my personal email delivered by UUCP. Seeing FreeBee is a blast from the past to be sure.
Back in the early 90's, I had a friend with one of these. It was a neat machine. I remember an unusual windowing system on it ("MGR", I think?) We were starting to get into PC unix clones - first Coherent, then Linux (SLS distro?) - around the same time.
MGR was a later one - it could run (slowly) with an unmodified machine but really needed a VIDPAL to be used best. VIDPAL was a replacement PAL which enabled user-mode access to the framebuffer.
As an intern during college I remember installing the OS on a 3B1 at one of my first summer jobs. So many floppy diskettes. Don't remember much more about it other than it seemed pretty primitive compared to the Sun workstations that were prevalent around the office.
Didn't SDF use this machine?
No a 3B2, which is different.
And probably the only real at-scale deployment of the Western Electric 32000 (Bellmac 32) processor. Sadly that chip wasn't very popular.
As LeoPanthera said above, Bell's exchange for the ability to make PCs was a bit of a bad swap in retrospect. It's pretty easy to see they were going for some kind of convergence of computing and telecoms (the 3B1 has an internal 1200 Baud modem) but they didn't do well on that bet.