Path: utzoo!utgpu!water!watmath!clyde!bellcore!tness7!killer!csccat!loci
From: loci@csccat.UUCP (Chuck Brunow)
Newsgroups: comp.binaries.ibm.pc.d
Subject: Re: Standard for file transmission
Message-ID: <629@csccat.UUCP>
Date: 18 May 88 06:36:52 GMT
References: <292@cullsj.UUCP> <696@fig.bbn.com> <2932@cognos.UUCP> <3055@encore.UUCP> <4776@teddy.UUCP> <3075@bsu-cs.UUCP>
Organization: Texas
Lines: 26

In article <3075@bsu-cs.UUCP> dhesi@bsu-cs.UUCP (Rahul Dhesi) writes:
>To avoid ambiguity, I suggest the following terminology.
>
>file in benchmarks, e.g. your UNIX kernel stripped of symbols, so there
>is some degree of consistency.
>-- 

	Using UNIX files as test cases misses the point. Unix sites
	rarely use binary as a mode of transfer. It would be reasonable
	to use MS-DOS files, as they represent the real thing.

	It is nearly universal that Unix sites use UUCP, in some form
	which will allow un-uue'd (pure binary) files. This is one
	reason why the question of compressing compressed files
	came up with in the first place: Text files can be batched
	and compressed to speed tranmission.

	Binary files are the fly in the ointment. Add the that the
	problem of handling assorted mailers for non-Unix sites
	and chaos begins. Embedded compression, (arc, gif, et al)
	complicate the situation even further.

	Numerous people have suggested FIDONET as a viable solution.
	Why not drop binaries from Unix sites and route them through
	FIDO? Then there is no problem of compressing compressed
	files.