This first paper is absolutely wrong.  When converting from a 32 bit system to a 64 
bit system you cannot just treat the 32 bit data as 64 bit data even if the systems 
are both the same endian.  I have experience converting from PDP-11 to VAX, and from 
Intel 286 to 386.  It just does not work the way the little endian proponents say it 
will.

The second is half right.  The addition of little endian mode was so that OS/2 could 
be ported to the Power architecture (actually PPC, which was the first to have little 
endian mode).  The problem was that OS/2 could not be ported to a big endian 
architecture.  I don't know why.  Note that Microsoft has stated that Windows/NT 
cannot be ported to big endian.  I suspect that the reasons are the same for OS/2, and 
NT.  Unix has been ported to both little endian, and big endian systems.  I suspect 
porting AIX to a little endian would be trivial.  There just isn't much motivation to 
do it.  With the death of OS/2, and the fact that OS/2 for PPC was never released, I 
doubt that IBM will bother.

-----Original Message-----
From: Ferguson, Neale [mailto:[EMAIL PROTECTED]
Sent: Wednesday, August 06, 2003 10:59 AM
To: [EMAIL PROTECTED]
Subject: Re: big and little endian


A couple of interesting references:
http://www.noveltheory.com/TechPapers/endian.asp
http://oss.software.ibm.com/pipermail/icu4c-support/2003-June/001664.html

Reply via email to