von Neumann documentary clip
hilpert at cs.ubc.ca
Thu Feb 26 00:19:15 CST 2015
On 2015-Feb-25, at 9:11 PM, Jon Elson wrote:
> On 02/25/2015 08:33 PM, Brent Hilpert wrote:
>> Brief clip from a 1966 documentary about von Neumann:
>> Goldstine presenting while standing in front of the IAS machine, which appears to be decommissioned and in storage.
>> Fuel for the eternal fire: Goldstine seems to be pretty definite in attributing the stored-program concept to Johnny.
> Well, I'm not sure the stored program concept was unique to von Neumann, it
> was one of those concepts that was entirely ripe at that time.
I was alluding to the never-ending dispute/question as to the origins of the stored-program concept.
I'm agnostic about it myself, as no one (now) really knows who said what in the meetings leading to von Neuman's production of "Draft Report on the EDVAC". He got the credit as his name was on the draft and it 'unintentionally' was distributed. IIRC from readings, vN himself didn't claim unique credit while Eckert/Mauchly claimed some degree of credit. On the other hand, (again IIRC) Goldstine was in those meetings, so would be a first-hand source.
> Stored program
> is meaningless without a decent storage component. Delay lines had a bunch
> of problems, Williams tubes were pretty limited, and drums were really slow,
> but you got more capacity. And, then, of course, core obsoleted all of the
> above, but that was well after the stored program concept was reality.
> But, something that still bears von Neumann's name is the concept of a
> unified memory for both data and instructions.
Yes . . which many consider a misattribution.
There is another interpretation of "von Neumann architecture": that of the IAS machine which went bit-parallel, breaking from the more prevalent bit-serial design of the first S-P machines, to become the more common form, and which might be said to be adding to the confusion over the designation.
> The computer in the video clip is the IAS, the cylinders along the bottom
> are Williams tubes, there's a row down each side of the machine. It is
> on display at the Smithsonian.
Found the full doc, indeed it mentions the location shown as a Smithsonian warehouse.
I didn't know or had forgotten it was at the Smith.
More information about the cctalk