The birth of random-access memory

Pres Eckert independently discovered the stored program concept, which von Neumann included in his EDVAC report. Now we call it a “von Neumann architecture” when it really should be “Eckert architecture,” or one of the British independent discoverers. As you said — von Neumann doesn’t need that recognition — and he likely used his position as a revered wartime physicist to grab it for himself anyway.

This article is not about of the invention of the computer with stored program, but about the invention of the random-access memory.

The Williams cathode-ray tube was the first kind of random-access memory without moving parts and the first widely known and widely used kind of RAM for computers.

I do not remember whether any of the earlier electromechanical computers with relays had any register set made with relays that could be considered as a RAM.

While there has been a report written by John Atanasoff in 1940, which has a very good discussion of all the kinds of elementary memory devices that can be used to store a digit in a computer and which describes the implementation of a dynamic memory with capacitors, which needs refreshing like the later Williams tubes or the modern semiconductor DRAM, but which has sequential access, the report written by John von Neumann in 1945 is the first public document that discusses the differences and the advantages or disadvantages of the memories with sequential access and of the memories with random access.

For memories with random access, John von Neumann identified as the most promising technology in the near term the cathode-ray tubes. He discussed a lot the possibility of using the already existing iconoscope tubes (used in video cameras) and he concluded that some improvements are needed in order to use CRT's as memories.

Like for the many other research teams which began to work around that time with the purpose of making an electronic digital computer, the work of the Manchester team was started by the John von Neumann report.

They had the merit that, unlike the other teams, they did not choose the easier path of using delay lines as memories, but instead of that they have chosen to follow the suggestion of John von Neumann and develop a random-access memory based on a cathode-ray tube, and eventually they have succeeded to do that. The reason why they have made this choice was that they had plenty of experience in making designs with CRT's for the British radars, during WWII.

For a few years, all the best computers in the world have used Williams tubes, e.g. the first generation of commercial IBM electronic computers, until the magnetic core memories completely replaced them.

So the merits for the first RAM used in computers are divided between John von Neumann, who has very clearly explained in 1945 why a RAM made with a CRT would be advantageous and what must be done to be able to make such a memory, and the Manchester team, mainly Frederic C. Williams and Tom Kilburn, who have then succeeded to design and build (and patent) a practical RAM using a CRT, during 1946 to 1948.

Regarding the John von Neumann report from 1945, it is difficult to distinguish which of the included ideas come from von Neumann himself and which have been learned by him during discussions with members of the ENIAC team.

In the case of the random-access memory, including its implementation with a CRT, that is an idea that seems to belong to von Neumann, because the members of the ENIAC team (i.e. Presper Eckert and John Mauchly) have based all their later projects, up to the first UNIVAC I in 1951, only on memories with sequential access.

In any case even for the ideas from the von Neumann report that might have originated from the ENIAC team, I am much more grateful to John von Neumann, who has explained them much more clearly than any member of the ENIAC team was ever able to do, and who let them become a public knowledge, which started immediately a very large number of research projects, and which lead to the creation of the computer industry.

The ENIAC team was secretive, because they hoped to start their own private computer company and become rich. I do not care about their dreams of becoming rich and in the following years they succeeded eventually to make the first commercially available computer in USA (UNIVAC I), but their computers had only very modest technical innovations over what was already presented in the von Neumann report.

All the major improvements in computer technology, e.g. the RAM, have been done in other research projects, which would not have existed without the publication of the John von Neumann report. Without that report, the progress of the computer industry might have been delayed by up to a decade.

While the British have made a lot of advances during the WWII in the theory and design of digital circuits made with electronic tubes (for using them in radars and decryption machines), they did not make any progress in the design of general-purpose computers. The German computers of Zuse were much more advanced.

Their experience in digital electronics from WWII enabled the British to have 3 teams (one of them including Alan Turing, who had very innovative ideas about extra instructions to be included in the ISA) and all 3 teams succeeded quickly to develop digital electronic computers, ahead of many American teams who began similar efforts at about the same time.

However the work of the British teams, exactly like the work of the American teams, has also been prompted by the publication of the John von Neumann report, and not by any experience from the war (which helped only in the practical implementation of the computing circuits).

So even if many ideas from the John von Neumann report do not originate in him, the publication of this report had a much larger practical impact than any earlier sources and it cannot be said that its recognition is too great.

The fact that the British have made in several cases some important discoveries but they have chosen to keep them secret is an interesting historical fact, but that gives them no merit whatsoever. For example, it does not matter if some British have discovered public-key cryptography in 1970, the only people who have merits about this are Ralph Merkle, Whitfield Diffie and Martin Hellman, who have made known this method to all humans in 1975.

Yes. Wikipedia actually has more information than the IEEE article. https://en.wikipedia.org/wiki/Williams_tube#Working_principl...