home *** CD-ROM | disk | FTP | other *** search
- Path: deepthought.xyplex.com!milan
- From: milan@xyplex.com (Milan J. Merhar)
- Newsgroups: comp.arch,comp.lang.c
- Subject: Re: Did Microsoft decree a byte order?
- Date: Thu, 4 Jan 1996 17:00:20
- Organization: Xyplex Inc.
- Message-ID: <milan.271.0C1B5523@xyplex.com>
- References: <4c80dp$9du@hobbes.sco.COM> <DKIp84.9Az@calcite.rhyolite.com> <jgkDKMn2x.2KA@netcom.com> <DKMx16.E2M@calcite.rhyolite.com>
- NNTP-Posting-Host: deepthought.xyplex.com
- Keywords: byte
- X-Newsreader: Trumpet for Windows [Version 1.0 Rev B]
-
- The only common disk or tape format that I am aware of that is rigorously
- byte-order-neutral is ISO9660, which goes to the extremes of recording
- EVERY value TWICE (once each way), so that no reading application ever has to
- do an order translation.
-
- This only works because the format was intended for a read-only medium (even
- though it is now used for read-mostly media too) so the overhead of
- calculating reversed-order values at write time could be justified by a great
- simplification of read-time processing. (Note that the reading application
- DISREGARDS the "inappropriate" data.)
-
- Given this exception (and its rationale for existing), Vernon is correct
- in stating that the best "real world" behavior is to pick a byte order and
- live with it, even if some number of applications will need to do order
- reversal. Agonize over which way to go, calculate which way minimizes the
- grief, but then freeze the result and live with it!
-
-
- Milan J. Merhar milan@xyplex.com
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-