[Info-vax] HP wins Oracle Itanium case

glen herrmannsfeldt gah at ugcs.caltech.edu
Mon Aug 27 17:25:55 EDT 2012


Bob Koehler <koehler at eisner.nospam.encompasserve.org> wrote:
> In article <BbPZr.13452$g43.7222 at fx21.am4>, ChrisQ <meru at devnull.com> writes:
>> 
>> Give me a few common use examples of real life computing where a byte means
>> anything other than 8 bits. I'm sure there must be a few.

>   TOPS-10 and -20 stored all thier ASCII characters in 7 bit bytes,
>   5 to the 36 bit word, with the msb of the word unused.

I believe LSB, as it fills from the MSB down. (Big endian.)

>   The COBOL compilers for TOPS-10 and -20 would store data in EBCDIC,
>   if asked, instead of ASCII.

I don't remember ever running it, but COBOL.SAV was my favorite
large file to test tapes with. Write a few copies on, read them
back of, see what happens.

>   EBCDIC was stored in 9 bit bytes, 4 to the 36 bit word, with the msb
>   of each byte unused.

I remember it was possible to either read or write all eight bits
of nine track tape, but don't remember how to do it. Otherwise,
as ASCII data the high bit would be ignored. 

>   So all the real life computing we did on TOPS-10 and -20 were
>   examples where a byte means something other than 8 bits.

I used to read and write tapes from IBM S/360.

>   The PDP-10 hardware reference defined "byte" as a user defined quantity
>   of 1 to 36 bits in length.

-- glen



More information about the Info-vax mailing list