Home page logo
/

interesting-people logo Interesting People mailing list archives

Re: 1620 One of your readers...
From: David Farber <dave () farber net>
Date: Wed, 3 Sep 2008 15:17:52 -0400



Begin forwarded message:

From: George Sadowsky <george.sadowsky () attglobal net>
Date: September 3, 2008 1:31:55 PM EDT
To: dave () farber net
Subject: [IP] Re:    One of your readers...

Dave,

For IP, if you like.

The IBM 1620 was nicknamed the CADET = Can't Add, Doesn't Even Try.

As I recall, the add table was contained in locations (single digits) 100-199, so when the computer wanted to know what was the value of x +y, it simply looked in location 1xy. The add table could be reloaded by any user program, so that it was easy to do arithmetic calculations in number systems with any base less than 10 also. Student hackers at the time found a number of other interesting ways in which a manipulated add table could be used.

The 1620 was also the first machine on which one could do serious variable length arithmetic, since it had a maximum storage capacity of 60KB (The IBM 1401 at that time was limited to 16KB). Using (I believe) a Taylor's series approximation for the value of pi, one weekend at Yale we compute pi to almost 20,000 decimal places by writing a short program, defining three very large numeric field, and letting the computer run all weekend.

And it was _possible_ to let it run all weekend because there were no tubes an the transistors did not fail!

George Sadowsky

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

At 11:31 AM -0400 9/3/08, David Farber wrote:
Begin forwarded message:

From: "Bob Frankston" <Bob19-0501 () bobf frankston com>
Date: September 3, 2008 10:17:59 AM EDT
To: <dave () farber net>, "'ip'" <ip () v2 listbox com>
Subject: RE: [IP] Re: One of your readers...

In the mid 80's I happened to notice a book on the shelves at local bookstore titled "Electronic Computers" by Henry Jacobowitz. Turns out it was a 1963 book with transistors being a new technology. About halfway through it says "Transistors have largely replaced tubes in computers because they need only a few volts for operation and do not require any heater power at all".

$7 (1960) may seem like a lot but according to http://en.wikipedia.org/wiki/IBM_7094 an IBM 7090 cost $2,900,000 in 1960 prices.

Considering that prodigious power consumption, reliability issues (which could be improved by keeping the tubes always powered up) and the ability to cover the high costs the decision to move to transistors would seem to have been a conservative decision with the risks being those associated with any developing technology.

Would a machine like the IBM 1620 even have been feasible with tubes? Of course the first model of the 1620 couldn't afford to waste transistors on luxuries like an adder so it relied on memory lookup instead.


From: David Farber [mailto:dave () farber net]
Sent: Wednesday, September 03, 2008 06:40
To: ip
Subject: [IP] Re: One of your readers...



Begin forwarded message:

From: "Rebecca S. Taylor" <becky () td-partners com>
Date: September 2, 2008 3:18:21 PM EDT
To: dave () farber net, spaf () cerias purdue edu
Cc: ddeininger () td-partners com, becky () td-partners com
Subject: Re: [IP] One of your readers...
Reply-To: becky () td-partners com

 Gene,

my business partner, Dick Deininger, worked at IBM from 1955-1984.
He sat in the circa Q3 1959 meeting where Vin Learson (then President of IBM) announced that "all future IBM computing products would be made using transistor logic". Vin made it clear that anyone who did not want to participate in this major transition was free to leave the company. The meeting lasted all of five minutes, after which the attendees were told to "get back to work".

He believes your figures are accurate, likely within a few pennies.

Dick is a treasure trove of information like this, feel free to call us if you like.

Kind regards,
Rebecca S. Taylor
General Partner
Taylor-Deininger Partners
Austin, Texas
www.td-partners.com

(512) 288-7352 (o)
(512) 657-1066 (m)



-----Original Message-----
From: David Farber [mailto:dave () farber net]
Sent: Tuesday, September 2, 2008 03:11 PM
To: 'ip'
Subject: [IP] One of your readers...

Begin forwarded message: From: Gene Spafford Date: September 2, 2008 1:39:07 PM EDT To: David Farber Subject: One of your readers... Dave, Perhaps one of your readers would know the answer to this. I've tried various online and book searches and cannot find the answer. Some time back, I read (somewhere) that Thomas Watson, Jr. took a big gamble in going from the IBM 700 series to the 7000 series by using all transistor logic. This was circa 1959. The figure I recall is that transistors were about $7 each (in 1959 dollars; about $60 each in today's dollars). I'm looking for confirmation and/or a source citation for this. Thanks in advance. -- spaf ------------------------------------------- Archives: https://www.listbox.com/member/archive/247/=now RSS Feed: https://www.listbox.com/member/archive/rss/247/ Powered by Listbox: http://www.listbox.com

Archives






-------------------------------------------
Archives: https://www.listbox.com/member/archive/247/=now
RSS Feed: https://www.listbox.com/member/archive/rss/247/
Powered by Listbox: http://www.listbox.com



Begin forwarded message:

From: "Bob Frankston" <Bob19-0501 () bobf frankston com>
Date: September 3, 2008 10:17:59 AM EDT
To: <dave () farber net>, "'ip'" <ip () v2 listbox com>
Subject: RE: [IP] Re: One of your readers...

In the mid 80's I happened to notice a book on the shelves at local bookstore titled "Electronic Computers" by Henry Jacobowitz. Turns out it was a 1963 book with transistors being a new technology. About halfway through it says "Transistors have largely replaced tubes in computers because they need only a few volts for operation and do not require any heater power at all".

$7 (1960) may seem like a lot but according to http://en.wikipedia.org/wiki/IBM_7094 an IBM 7090 cost $2,900,000 in 1960 prices.

Considering that prodigious power consumption, reliability issues (which could be improved by keeping the tubes always powered up) and the ability to cover the high costs the decision to move to transistors would seem to have been a conservative decision with the risks being those associated with any developing technology.

Would a machine like the IBM 1620 even have been feasible with tubes? Of course the first model of the 1620 couldn't afford to waste transistors on luxuries like an adder so it relied on memory lookup instead.





-------------------------------------------
Archives: https://www.listbox.com/member/archive/247/=now
RSS Feed: https://www.listbox.com/member/archive/rss/247/
Powered by Listbox: http://www.listbox.com

  By Date           By Thread  

Current thread:
  • Re: 1620 One of your readers... David Farber (Sep 03)
[ Nmap | Sec Tools | Mailing Lists | Site News | About/Contact | Advertising | Privacy ]