Path: utzoo!utgpu!watserv1!watmath!att!rutgers!usc!samsung!munnari.oz.au!murtoa.cs.mu.oz.au!ditmela!yarra!bohra!ejp
From: ejp@bohra.cpg.oz (Esmond Pitt)
Newsgroups: comp.std.misc
Subject: Re: Binary data formats
Keywords: file formats, binary, integer, floating-point
Message-ID: <221@bohra.cpg.oz>
Date: 4 Dec 89 06:54:52 GMT
References: <4021@hydra.gatech.EDU>
Reply-To: ejp@bohra.cpg.oz (Esmond Pitt)
Organization: Computer Power Group, Melb, Australia
Lines: 31

In article <4021@hydra.gatech.EDU> gg10@prism.gatech.EDU (Galloway, Greg) writes:
>
>Problem:
>
>* Integer data (to my knowledge) only consists of two types:
>
>  - Big-endian, or Intel-type: Most significant byte first (16 or 32 bit)
>  - Little-endian, or Motorola-type: Least significant byte first

(A) This seems to be back to front. LSB-first is big-endian, from the
point of view of memory addresses.

(B) I may be the 10,000th person to state this, but there is also the
issue of 'short' order within a 'long', which is either {LSW,MSW} or
{MSW,LSW}.  Permute this with the byte order and you get 4
combinations, at least 3 of which (I believe) are seen in practice:

		{LSW{LSB,MSB},MSW{LSB,MSB}}	/* 'big-endian'	*/
		{MSW{LSB,MSB},LSW{LSB,MSB}}	/* no instances known to me	*/
		{LSW{MSB,LSB},MSW{MSB,LSB}}	/* PDP-11		*/
		{MSW{MSB,LSB},LSW{MSB,LSB}}	/* VAX, Intel	*/

For 64-bit data there are 8 possibilities, and so on ...

-- 
Esmond Pitt, Computer Power Group
ejp@bohra.cpg.oz

-- 
Esmond Pitt, Computer Power Group
ejp@bohra.cpg.oz