Face down, nine-edge first (wherein all is revealed)

Alan Kay Alan.Kay at disney.com
Tue May 9 21:31:13 UTC 2000


The Computer Museum (at Moffet Field near San Jose) has many old machines,
and is currently restoring a 1620.

Cheers,

Alan

-----

At 12:23 PM -0800 5/9/00, David L. Caster wrote:
>Ed,
>
>> Well, maybe a 1620 port - someone in the Pacific Northwest has one that
>> apparently works.  This machine would do a great job with LargeIntegers,
>since
>> the word length was limited only by the size of the core memory (20 to
>60k,
>> depending on the depth of your pockets) and the fact that you needed room
>for
>> both operands (and possibly the program, too).  You can even redefine
>arithmetic,
>> which was table-driven!  The Model 2 had an adder, but the Model 1 used
>both
>> multiplication and addition tables.  You could get disk drives for these
>> machines, and the two-pass FORTRAN-IID (D for disk) compiler was actually
>pretty
>> good.  The 1620 was supposed to have been the "scientific" counterpart to
>the
>> "business" 1401.  The 1620 was my introduction to computers (1968) - does
>anyone
>> else remember this wonderful machine?
>
>Absolutely.  In fact, I can still tell you the cold start sequence for that
>machine with Monitor II: 3400032007013600032007024902402111963611300102.
>Must have typed that a million times.
>
>In the OBTW department, Multiply and divide was limited to 100 digits.  Add
>and subtract operands could be anywhere in memory and as you said could be
>as big as memory.  No registers either, at least as far as the programmer
>was concerned--none of this LDA STA phony-baloney!  Actually, quite a
>wonderful machine, one of which I would like to have.  I understand that
>the last one was delivered by IBM to the Smithsonian.
>
>We really should remove this reminiscing to a list devoted to computer
>history and archeology, though.  [8~{)>  One the other hand, one of my
>colleagues finds it reassuring that there are few of us old hands working
>on this bleeding edge stuff.
>
>David L. Caster







More information about the Squeak-dev mailing list