Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!caip!topaz!husc6!seismo!rochester!kodak!elmgate!jdg
From: jdg@elmgate.UUCP (Jeff Gortatowsky)
Newsgroups: net.micro.amiga
Subject: Re: Orphaned Response
Message-ID: <494@elmgate.UUCP>
Date: Wed, 9-Jul-86 15:56:59 EDT
Article-I.D.: elmgate.494
Posted: Wed Jul  9 15:56:59 1986
Date-Received: Fri, 11-Jul-86 06:06:31 EDT
References: <8606161842.AA20107@pavepaws> <26600033@ccvaxa>
Organization: Eastman Kodak Co., Rochester, NY
Lines: 52
Summary: Size of shorts int and longs

In article <26600033@ccvaxa>, aglew@ccvaxa.UUCP writes:
> 
> I've noticed a lot of people talking about shorts and ints and longs,
> and whether an int should be 16 or 32 bits, and what the size of a 
> short should be.
> 
> I've also noticed somebody say that he always uses shorts or longs,
> because they always have a definite size. SHAME! Where is it written
If it's my message you refer to I guess:
  A: I mis-stated what I meant. On the Amiga shorts are 16 bit and longs
are 32 (yes It's not cast in concrete) but it makes sense and _WORKS_.
  B: As I state below this assumption is QUITE portable to larger data size
machines.
  C: When I choose a data type I choose the smallest that will hold the
expected value of the data. That means shorts for data <65536.  It's
faster and portable (Blast you megamax!) to 16 bit CPUS or >.
> in stone that shorts are 16 bits and longs are 32 bits, and ints one
> or the other? What if you have a machine that supports 64 bit integers
> in hardware - wouldn't it be natural to have shorts 16 bits, ints 32 bits,
> and longs 64 bits? Sure it would be! - but you can't do the sensible
> thing because of all sorts of brain-damaged non-portable programs that
> assume an int is the same thing as a long, 32 bits worth. Which leads
> to abominations like long longs, etc.
> 
> This is what is going to kill C - not Ada, just the accumulation of
> implicit machine dependencies.
> 
> Andy "Krazy" Glew. Gould CSD-Urbana.    USEnet:  ihnp4!uiucdcs!ccvaxa!aglew
> 1101 E. University, Urbana, IL 61801    ARPAnet: aglew@gswd-vms


I don't know if it was me your refering to as using longs and shorts only,
but thats about what I do.  Your point about 64 bit machines doesn't apply.
I don't assume anything except that ints are non-portable, period. If
shorts are 16 or 32 bits on a 64 bit machine, my code works. Further on a
restricted bus machines (like 68000 and 68010 or a 64 bit CPU on a 32 bit
bus) my coding practice is a big win. If longs are 32 bits or 64 bits my
code works. Can you say the same for ints? 

I think we are on the same side.  I don't assume anything about the SIZE of
a data type. But I do assume shorts and longs will be upward (size wise)
compatible. 

BTW Longs would(should) be 64 bits, not long longs (gimme a break 8-)). 
double long?
Nah....


-- 
Jeff Gortatowsky       {allegra,seismo}!rochester!kodak!elmgate!jdg
Eastman Kodak Company  
<Kodak won't be responsible for the above comments, only those below>