Subj : Re: Computers
To   : Spectre
From : tenser
Date : Tue Jul 23 2024 03:13:39

On 22 Jul 2024 at 08:21p, Spectre pondered and said...
 
 Sp> While I agree with your appraisal, effective 2:1 compression seemed to be
 Sp> pretty rare.  Binaries and hard data didn't tend to compress as well as a
 Sp> chunk of equivalent sized text.  I don't remember getting beyond ~40%
 Sp> compression as a general rule, and small files didn't tend to compress as
 Sp> well as larger ones.

Yeah.  It depends very much on _what_ you're compressing
and how much overhead you're willing to tolerate.  Still,
even if you only get 30%, you're about breaking even with
`uuencode`.

 Sp> Its kind of weird in a way, a lot of stuff was optimised for slow speed
 Sp> transmission, 2400bps or so.. so things tended to be fairly small, then
 Sp> you'd compress it to get it down further... worked point to point out of
 Sp> a BBS nicely, but then you go and add overhead to get through something
 Sp> like a 7bit mail server... signs of the times I guess.  It was always
 Sp> the last link... ISP or Uni or whatever it was you connected to, to
 Sp> home.. the weakest, slowest link.

Yeah.  People tend to forget that we had to contend with
7 bit systems, fixed records sizes, and all sorts of other
weird things that made interoperability hard.  Our IBM
mainframe running VM/CMS used to eat the ends of lines that
were longer than 80 columns when transferring email (which
all had to be 7-bit clean, of course, as it would be
translated between ASCII and EBCDIC).

--- Mystic BBS v1.12 A48 (Linux/64)
 * Origin: Agency BBS | Dunedin, New Zealand | agency.bbs.nz (21:1/101)

.