I use the following mental model of the "generations" of early electronic computers (the mechanical ones are a whole hierarchy unto themselves):

• Zeroth: Electronic computing devices which were not true stored-program computers - Colossus computer (misnamed article), ENIAC
• First: One-off tube computers: Pilot ACE, Manchester Mark 1, EDSAC, SEAC, EDVAC, IAS, Whirlwind, etc
• Second: Early production tube machines: Ferranti Mark 1, DEUCE, UNIVAC I, ERA 1101, IBM 701, IBM 650, IBM 702
• Third: Mature production tube machines: IBM 704, IBM 705, AN/FSQ-7
• Fourth: Experimental transistor machines: TX-0
• Fifth: Early production transistor computers: PDP-1, IBM 7090
• Sixth: Mature transistorized machines: PDP-6, KA-10
• Seventh: Experimental IC machines: MU-5
• Eighth: Early production IC computers: PDP-8, PDP-11/20

Anything after that can't possibly be called an "early" computer!

Note: Machines are listed for illustrative purposes only - this list is not by any means intended to be a comprehensive list of all the machines of any particular generation.

• Note that these categories of early computers conflict with what are generally called the computer generations: 1 = tubes, 2 = transistors, 3 = integrated circuits, 4 = microchips. I think they should be called something different to avoid confusion. Something such as zeroth=1A, first=1B, second=1C, third=1D, fourth=2A, etc.

--Bubba73 18:53, 26 May 2005 (UTC) Also, the Harvard Mark I and IBM SSEC sould probably be included as your "zeroth", changing the ones following, making the first all-electronic ones the "first". Bubba73 July 6, 2005 15:11 (UTC)[reply]

I'd suggest that the Colossus computer article is not mis-named, because the Colossi were definitely digital (partly) electronic computing machines. They were also programmable to a limited extent. What they were not was either general purpose or Turing-complete.

As with the original version of ENIAC, Colossus was not software programmable.

86.141.196.130 (talk) 23:10, 5 January 2014 (UTC)[reply]