home *** CD-ROM | disk | FTP | other *** search
- Path: senator-bedfellow.mit.edu!jfc
- From: jfc@mit.edu (John Carr)
- Newsgroups: comp.arch,comp.lang.c
- Subject: Re: Did Microsoft decree a byte order?
- Date: 4 Jan 1996 15:20:38 GMT
- Organization: Massachusetts Institute Of Technology
- Message-ID: <4cgr86$6hd@senator-bedfellow.MIT.EDU>
- References: <4b56do$c3u@sundog.tiac.net> <DKIp84.9Az@calcite.rhyolite.com> <jgkDKMn2x.2KA@netcom.com> <JAN.96Jan4144902@cora.neuroinformatik.ruhr-uni-bochum.de>
- NNTP-Posting-Host: kangaroo.mit.edu
-
- In article <JAN.96Jan4144902@cora.neuroinformatik.ruhr-uni-bochum.de>,
- Jan Vorbrueggen <jan@neuroinformatik.ruhr-uni-bochum.de> wrote:
-
- > What if GIF files didn't work between machines? Or what about tar files?
- >
- >Irrelevant. These formats are expressly designed for interchange, and they
- >carry a hefty performance penalty because of it. The X protocol is another
- >example. For almost all other cases, the equivalent of a core dump is just
- >fine.
-
- The X protocol allows the client to choose a byte order, and the
- standard X library sends binary structures over the net (it took
- a while to get this right on the Cray, but C compilers on most
- systems make it easy to create a data structure that corresponds
- to the X protocol byte stream). The server swaps bytes only if
- it has a different byte order. Usually this is not the case (the
- server is on the same machine or both client and server are the
- same type of hardware).
-
- --
- John Carr (jfc@mit.edu)
-