[Info-vax] HP wins Oracle Itanium case

[Johnny Billquist]

On 2012-08-24 19:52, ChrisQ wrote:
> On 08/24/12 15:47, Johnny Billquist wrote:
>
>> You are (again) in the text assuming that a byte is 8 bits.
>
>> Having anything use 8-bit bytes on a PDP-8 is extremely inefficient.
>
> Nitpicking. Most people understand a byte to mean 8 bits, where as a
> pdp8 had
> a 12 bit wordsize. The byte efficieny, or not, of that machine is not
> relevant
> here.

It's not nitpicking. You made a comment to something I wrote, and what I 
wrote was literally: "However, a byte does not necessarily imply 8 bits... "

My whole point was exactly what you demonstrate here - a byte is not 
necessarily 8 bits, and have never been. Some people, however, makes 
that assumption.

> Using the term octet makes it even clearer.

Using the word octet is unambiguous, which is why the term even was created.

> 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.

Sure, how about the instruction BSW on a PDP-8. It reads out as "Byte 
SWap". But it swaps two 6-bit values.
Or how about a PDP-10, where the byte size is variable, and can be 
anything from 1 bit to 36 bits?

> Don't you just love the level of debate in this group :-) ?...

I'm sometimes enjoying it. Other times I just try to correct silly small 
things that irritate me and I spot, but which are really not important.
The case of byte size is almost like that. It don't really make much of 
any difference today, since almost all architectures nowadays only deal 
with 8-bit quantities anyway.

But like I said somewhere else before, lookup "byte" on wikipedia. 
Wikipedia is not always a good source of correct information, but it's 
not too bad in this case. :-)

	Johnny