You don’t have conversations with microprocessors. You tell them what to do, then helplessly watch the disaster when they take you literally!
David Brin
It was during the Thanksgiving holiday of 1969 when the engineer Victor Dale Poor (1933-2012) and his fellow amateur radio colleague Harry Starck Pyle (1949-2013) produced the underlying architecture of the modern microprocessor on a living room floor. Several months ago, Pyle, still a student, was hired for a computer-related summer job by his mentor Poor, who was the executive of Maryland-based radio and telegraph equipment manufacturer Frederick Electronics, where developed the idea of adapting radioteletype (RTTY) machines to send data wirelessly.
As Poor and Pyle came up with the idea of a high-density integrated circuit that would be programmable, they offered this circuit design to various IC manufacturers but were turned down by all of them. The reason? The chip was too specialized and would never have enough widespread applicability to be financially worth developing.
Having failed to convince any IC manufacturer, but still believing in the concept Poor and Pyle pressed on and went looking for manufacturers who would have an application for the new chip. They found such a firm Computer Terminal Corporation (CTC), which made a variety of lower-cost computer terminals which were compatible with various computer companies and was involved also in the development of the first microprocessors in another way. CTC was interested in the chip because it presented a way to make one terminal that could be programmed to behave like and handle the protocols of a variety of different computer manufacturers. CTC agreed to pay two different chip makers (Intel and Texas Instruments) to produce the chip. Both chip makers decided to use a new technology, called PMOS, to produce the chip.
Both chip makers have encountered new technology problems with PMOS and have failed to meet the CTC deadline. So the contracts are canceled and CTC builds the chip via discrete TTL logic and puts it into a programmable desktop terminal, called the Datapoint 2200. The original instruction set architecture was developed by Poor and Pyle. The TTL design they ended up using was made by Gary Asbell. Jonathan Schmidt wrote the accompanying communications software.
Datapoint 2200 was unveiled at the beginning of 1970 and was launched to the market in June. It becomes an extremely successful product and was manufactured as long as until 1979. So much so, that CTC changes its name to Datapoint Corporation. They go on to invent another new technology to connect all of their low-cost computers together. While the initial Datapoint 2200 did not have a microprocessor in it, it had the programmable equivalent of an Intel 8008 (built by discrete TTL logic) and it funded the initial development of the first microprocessor.
The Datapoint 2200 was a box with size 9 5/8” in height, 18 1/2” wide, 19 5/8” deep, and weight 47 pounds. It had a small built-in 7” x 3,5” CRT screen, which was of green&black monochrome type, and worked in text mode 12 x 80. It had full stroke keyboard+numeric keypad with an integrated programmable beeper. The external data and program storage were two read-write cassette decks for 130KB of mass storage. The internal memory of the first models was from 2 KB to 16 KB max. There was a run light and two other lights on the keyboard. The I/O ports were—RS 232, LAN connector, and printer (parallel data) connector.
The operating system was Datapoint O/S (cassette/storage drum-based O/S). When the machine halted, you could not tell where in the program it had done so. So you put in the O/S cassette and rebooted. You then did a memory dump and tried to deduce what had gone wrong. Primitive by today’s standards, it was the first computer on a desktop.
There was also an optional disk drive using Shugart 8″ floppies, single-sided, single-density. It was the first commercial computer to include them! A number of peripherals can be connected also: printer, tape device, 2 MB removable cartridge disk, card reader, etc. (see the Reference Manual of Datapoint).
Languages included Databus and Datashare (“COBOL-like” business computer languages, interpreted to allow multiple 80 character by 24-row dumb CRTs to share tiny partitions of RAM memory in the main system unit), a Basic interpreter, and an RPG II compiler (see the Programmers Manual of Datapoint).
The Datapoint 2200 was a real computer and it fits on a desk, but it certainly wasn’t priced for hobbyists, as it was sold for about $5000.