Subj : Re: bit shifting
To   : comp.programming
From : Rob Somers
Date : Sun Aug 14 2005 03:52 pm

> I don't think you've read Alex's post carefully enough. 

That is quite possible.

> By the looks of 
> it, you wrote your code wrongly because you thought a "short" was 8 bits
> when it is actually 16 bits. So your "~x" is not 127.

I wrote my code to simulate the conditions needed to create the same results
as the book I am using.  Were I to leave out the cast to short, I would
have generated 4294967167 from the ~128 operation.  The book (which is just
an old one I bought used) was apparently written at a time when an int
might have been 2 bytes.  

My apologies for not getting all of this too quickly - I do feel like a bit
of a dope.  Anyway, I *think* I understand, so lets see if what I think is
happening, is really happening.

On paper, if I write down the number 128(decimal) in binary, it would look
like this:  10000000

On paper, if I write down the number 127(decimal) in binary, it would look
like this: 1111111

On paper, if I take 10000000 and do a complement operation on it, I get
this:  01111111

But, on my computer,I am thinking 128 binary might look something more like
this:  000...010000000, so that when I perform a complement operation on
it, all the bits *before* the 1 and *after* get flipped, like so: 
111...101111111.  Does that make sense, and I am correct?

.