25 Jan
2012
25 Jan
'12
10:36 p.m.
On 2012 Jan 25, at 1:28 PM, Mike Loewen wrote:
On Wed, 25 Jan 2012, Brent Hilpert wrote:
Thanks for the references, some great reading, but I'm not clear what
your intention or position is. I don't see anything in those papers
to contradict my points, I believe they support them.
The Williams-tube symposium of the first paper was held in Dec 1951,
long (years) after the Whirlwind project had been working at
developing CRT-storage. The paper mentions other projects working on
Williams-tube implementations (SEAC, ORDVAC, Los Alamos,..), but does
not mention Whirlwind in that group. By that time Whirlwind was
struggling with the inadequacies of their CRT memory, my take on that
paper would be that Everett/Whirlwind was following the efforts of
projects using Williams-tube CRT storage, perhaps under consideration
of it being an alternative to their own (non-Williams) CRT storage.
The first paper states: "One of the early proposed advantages of the
Williams'-type storage, that of the relative economy of using
ordinary cathode-ray tubes, seems to be hard-pressed for
confirmation." Along with other discussion in that paper, along with
the paper detailing the Whirlwind design, one can see how different
the Williams-tube and Whirlwind designs were.
On 2012 Jan 24, at 12:20 PM, Mike Loewen wrote:
When in doubt, go to the source:
"Symposium on Williams'-Tube Storage"
http://dome.mit.edu/bitstream/handle/1721.3/39201/MC665_r05_M-1380.pdf
"Initial Testing of a Computer Electrostatic Storage System"
http://dome.mit.edu/bitstream/handle/1721.3/39822/MC665_r11_R-195.pdf
"Storage Tube Program, Project Whirlwind"
http://dome.mit.edu/bitstream/handle/1721.3/38916/MC665_r04_M-102.pdf
An early version of the Whirlwind computer also used Williams
Tubes:
"Initially Whirlwind used a modified form of the Williams tube.
An additional flood gun maintained the pattern of dots while a
writing gun was used to alter the pattern. Thirty two such tubes
were needed to provide the 4KBytes of storage that the Whirlwind
needed. Given a tube life of one month and cost $1000 the running
cost of the machine was very high, $1 per bit per month."
Whirlwind never used Williams-tube storage, at least as I
understand the definitions. I don't believe it is even
historically accurate to say ".. Whirlwind used a modified form of
the Williams tube".
Williams-tubes memory (or the preferred "Williams-Kilburn-tubes")
and the first memory system of Whirlwind were both electrostatic
CRT-based storage. The design approach of the Williams-Kilburn-
tube-memory was intended to enable the use of standard or off-the-
shelf CRTs, while the Whirlwind design used specialised holding-
beam CRTs.