XPost: alt.folklore.computers   
   From: sfuld@alumni.cmu.edu.invalid   
      
   On 8/7/2025 7:57 AM, Anton Ertl wrote:   
   > Peter Flass  writes:   
   > [IBM STRETCH bit-addressable]   
   >> I don't get why bit-addressability was a thing? Intel iAPX 432 had it,   
   >> too   
   >   
   > One might come to think that it's the signature of overambitious   
   > projects that eventually fail.   
      
   Interesting.  While it seems to be sufficient to predict the failure of   
   a project, it certainly isn't necessary.  So I think calling it a   
   signature is too extreme.   
      
      
   > However, in the case of the IBM STRETCH, I think there's a good   
   > excuse: If you go from word addressing to subunit addressing (not sure   
   > why Stretch went there, however; does a supercomputer need that?)   
      
   While perhaps not absolutely necessary, it is very useful. For example,   
   inputting the parameters for, and showing the results of a simulation in   
   human readable format,  And for a compiler.  While you could do all of   
   those things on another (different architecture) computer, and transfer   
   the results via say magnetic tape, that is pretty inconvenient and   
   increases the cost for that additional computer. And there is   
   interaction with the console.   
      
      
   , why   
   > stop at characters (especially given that character size at the time   
   > was still not settled)?  Why not continue down to bits?   
      
   According to Wikipedia   
      
   https://en.wikipedia.org/wiki/IBM_7030_Stretch#Data_formats   
      
   it supported both binary and decimal fixed point arithmetic (so it helps   
   to have four bit "characters", the floating point representation had a   
   four bit sign, and alphanumeric characters could be anywhere from 1-8   
   bits.  And as you say, 6 bit characters were common, especially for   
   scientific computers.   
      
      
   > The S/360 then found the compromise that conquered the world: Byte   
   > addressing with 8-bit bytes.   
      
   Yes, but several years later.   
      
   Another factor that may have contributed.  According to the same   
   Wikipedia article, the requirements for the system came from Edward   
   Teller then at Lawrence Livermore Labs, so there may have been some   
   classified requirement that led to bit addressability.   
      
      
   > Why iAPX432 went for bit addressing at a time when byte addressing and   
   > the 8-bit byte was firmly established, over ten years after the S/360   
   > and 5 years after the PDP-11 is a mystery, however.   
      
   Agreed.   
      
      
      
   --   
     - Stephen Fuld   
   (e-mail address disguised to prevent spam)   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)