It’s no secret that Andrew Tanenbaum has disposed of some shady characters that have attacked Linux with less than truthful motives. On a warm Sunday in June of 2015, he stopped by for a tour of the collection and to tell us where some of the bodies are buried.
In 2004, as the failed Longhorn OS was being purged from Redmond hard drives, Microsoft hired a D.C. firm to do a smear job on Linux, Torvalds and Open Source Software. It seems they wanted to send a message that Open Source was not all the intellectual freedom it was cracked up to be.
According to their hired gun, Torvalds had copied Linux from Andrew’s Minux. Open Source was just plain bad for the economy. Taxes and license fees built that.
In a famous Usenet discussion Tanenbaum defended Linux and Open Source. Linus Torvalds did not copy Linux from Minux. In fact he had made some mistakes with the monolithic design of the kernel that Tanenbaum said was the purest form of self invention.
Andrew told me, “the development of communications technology paralleled that of computers.” A member of the original UNIX team and frustrated licensee, he did a teaching version called Minux.
Like UNIX, the modern CPU was suggested and designed by engineers in the communications business . In 1969, Victor Poor, working on an improved control unit for the Datapoint 2200, contributed logic for the 1201 – Intel renamed the 8008.
In 1971 the Datapoint 1200 TTL chipset was faster than its single chip silicon counterparts . This posed a serious challenge to engineers designing early CPU’s. An ex-Intel employee was about to dominate the 8-bit market with an improved version of the 8080.
Edward Kleinschmidt is the only inventor we know that had his own lab in the 19th century and still lived to see the Apple II. He was 101 years old when he passed in 1977.
But how come it took so long that only he made the journey with his amazing tech? I would argue that the mathematical science developed to help win WWII and the science of the cold war developed to land on the moon actually slowed down the process of innovation.
Conrad Zuse laughed at the prospect of using vacuum tubes. Conrad Zuse invented his first Z1 computer using telephone relays and old movie film for storage. An example of the Simon 1 relay computer can be seen in the Computer History Museum.
After looking at nearly all tech from electric razors to industrial robots I find that communication companies have been the driving force behind innovations like Kleinschmidt’s teletype. The path was not a straight one with dead ends like the 30-ton mainframes, Viatron “microprocessor” and Cray 4 .
Sorry UNIVAC fans but the 30-ton 1105 (in our collection – 29.99 tons) was absurd compared to the ABC Computer and the Z1. If 40’s transistor tech was properly evaluated, computers would have never grown as large as the ENIAC.
Claude Shannon’s “mathematical theory of communication” wins the day for the the modern PC. He was first to expose the benefits of non-analog systems and use the term “bit”.
Here is our 100 year old Kleinschmidt teletype. Plutarch’s parallel tech history might include this CCS Teleswitcher store and forward message switcher from 1971. This British black box used the PDP-8 as a back-end processor. It only sold a few installations perhaps because of its $100K price.
Both are notable for their small compact design. Contrast that with the larger than necessary footprints of the UNIVAC 9200 and IBM 370.
Just when we thought collectibles coming out of Texas in 2014 couldn’t get any better comes this gold version of the TI Datamath calculator made in 1973. Held by a Dallas man for nearly 45 years and still sealed in original factory plastic wrap. TI made a practice of special Dallas distributions of the Datamath. This is one of the rarest and most perplexing of any TI calculator ever discovered. It has a serial number and label consistent with other examples in the online Datamath Museum.
TI pioneered the calculator on a chip TMS100 in 1971 and was awarded a patent for first single chip microprocessor although microcontroller is a better description of the design. As to kick sand in the face of Intel 4004 fans calculator firm Busicom replaced the Intel chipset with a single TI chip in subsequent versions of it’s business calculators.
Unfortunately this model has corrosive batteries that are also entombed in the original plastic wrap and will need to be addressed at some time. I want to open this on this 50th anniversary of the Datamath in 2022.
Datapoint 3300 gets it’s name because it was 100 times better than the Teletype model 33. In 1974 Vic Poor directed John Murphy to create a networking standard that was plug compatible with the first Datapoint Terminals.
This stunning Datapoint breadboard prototype was discovered in a box from San Antonio, TX where it all started and stopped for the pioneering company. In the pictures we plug it in to a Datapoint I/O port. Also included is a finished product – a MSI logic Datapoint 8-port coaxial hub.
IBM created the PALM processor in 1973 for use in the IBM 5100 portable. Reaching back to the 360 mainframe, IBM created a portable version of APL with 64K of RAM. Consumers would have to wait another 7 years to get their hands on 64K of RAM but scientists and engineers had it in the 5100. Our 5100 “Property of IBM” is a flawless fully working example. With a simple “util” command we did a directory listing on our demonstration tape. Complete with a full set of manuals the 5100 is one of the holy grails of collecting. A working example so rare several major collections are without one. The Computer History Museum in Mountain View and Eric Klein’s online collection are both sans 5100. The powers of the 5100 were so overwhelming at the time that conspiracy theorist John Titor claimed the machine contained hidden translating code derived from an alien intelligence.
Thomas Watson Jr. had stepped down from IBM in 1971 after increasing profits ten fold to 1 billion dollars in that year and the new team at IBM was not about to let the Texas upstarts get a foothold in the small computer business. Datapoint claimed the 2200 fit in the footprint of an IBM Selectric typewriter but IBM upped its game with a first ever computer portable.
I am ready to defend the Datapoint 1201 as the world’s first alphanumeric microprocessor. Although TI was first able to produce the actual chip in the TMX 1795, Intel management decided it was worth losing Datapoint as a customer over a less risky, simply packaged, 4004 calculator chip .
“Faced with the facts that the logic design of the 8008 was made by Datapoint and its initial chip implementation was covered by a Texas Instrument patent application, Intel conferred on 4004 the status of the first microprocessor.” – Dan Alroy chairman of the 1975 IEEE conference.
At the suggestion of Alroy, Datapoint’s Gus Roche was to attend the 1975 conference and give a speech about his contribution. On the eve of the convention Roche was killed in a car accident.
Young Bill Gates next to a Datapoint 1800 programmable, stand-alone machine with 60K user RAM, 4K system ROM, and an 8-bit CPU built from a TTL chipset. In the modern photo the 1800 is replaced by a dual Z80, Intertec Superbrain.
As far as chips go the TI implementation of the 1201 has recently been rediscovered by Ken Sheriff in his excellent blog post “The first forgotten CPU : TMX 1795″ Unless someone has a “1201” prototype 16-pin version of the 8008 around the 4004 is as close to the first CPU as collectors have been able to come on the Intel side.
While the 4004 was groundbreaking Intel’s management effectively sabotaged the 1201 forcing designers to bow to Datapoint specs. The specs fit Andy Grove’s 16-pin package design for memory and no one was about to challenge the company package.
Intel’s management was focused on memory chips and viewed the 4004 as just another MOS semiconductor. When management finally realized the potential of the CPU they needed to buy the rights of the 4004 back from Japanese calculator maker Busicom.
Mostek’s calculator on a chip soon replaced the 4004 in Busicom’s business line while TI went into the calculator business themselves with the TMS100.
The movie Jurassic Park would not be filmed for another 23 years but Intel was still a small company in 1970 walking among the dinosaurs of technology. Known as the BUNCH, these companies had near total control of the large computer market. Intel had little choice but to make a memory board for UNIVAC. Their first board product was added to the company’s 1970 price list.
This board pictured above is Intel’s first board product from 1970. It is a memory board for the UNIVAC! Intel Memory expansion board for UNIVAC mainframes. The company’s first board product.
It contains four channels through 4 long 3M made ribbon cables. After nearly 50 years these ribbon cables feel as new as they day they were made.
Programming Note: This article was heavily revised for my Google+ Datapoint page but preserved here in it’s original version.
My first blog concerns the very first Intel compatible CPU board from a contract Intel signed in 1969 called the 1201. This is the TTL implementation done for the world’s first microcomputer the Datapoint 2200. It has four early production 3101 memory chips that provide an interconnect for external memory cards. Looks a lot like a PC motherboard to me!
Recently arrived in the lab is a Datapoint CPU board with 6947 date code on four blue memory slots each flanked by an Intel 3101 static memory chip – Intel’s first product from 1969. Fleetwood Mac was just climbing the charts when these babies were made.
In the first year of Intel’s operation memory chips were a priority over microprocessor design. The 3101 you see in the picture is Intel’s very first product having gotten done 8 months after the company’s founding. Intel then contracted with Computer Terminal Corp. to design a microprocessor called the 1201.
Two ex-NASA engineers at Computer Terminal Corp were fresh off a success with their 3300 terminal and wanted something with an “improved control unit” according to their business plan.
Enter Intel and Texas Instruments pitted in a battle to develop the world’s first 8-bit microprocessor. Both built the 1201 CPU for the Texans but in one of the biggest business blunders in history CTC did not secure the rights to the CPU for a mere fifty thousand dollars. They could have easily owned Intel at that point.
Instead the company rushed out our board in the pictures using TTL (transistor to transistor logic). It was actually quite fast. But thanks to the design’s inability to access memory directly through the blue “69” slots we have this stunning gold 3101 four-banger. The first x86 compatible in history.
Intel sent its main salesperson out into the field in the summer of 1971. He reported back that there was some tepid interest in the 1201 which Intel renamed 8008 to fit with the entirely incompatible 4004. But more on that later.