NetNews Usenet Archive 1992 #18/NN_1992_18.iso/spool/comp/sys/3b1/3063

Internet Message Format | 1992-08-12 | 1.7 KB

Path: sparky!uunet!dtix!darwin.sura.net!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!linac!att!pacbell.com!tandem!zorch!hico2!kak
From: [e-mail] (Kris A. Kugel)
Newsgroups: comp.sys.3b1
Subject: LAPM and uucp "f" protocol (was: Re: How to run 19,200 (or faster))
Summary: Will LAPM replace the need for error correction by "g" protocol?
Message-ID: <[e-mail]>
Date: 11 Aug 92 17:28:35 GMT
References: <[e-mail]>
Organization: High Country Software
Lines: 25

Given that the "LAPM" is supposed to supply an "error-correcting" link,
and using hardware flow control, is it possible that we will only
need the "g" protocol for non-V.42 connections?  This leads to some
distinct possibilities, but I would not be sure how to get my modem/uucp
system to make the distinction between V.42 and non-V.42.

: >However, "f" protocol is a big loss on large incoming files - it doesn't
: >work much at all.  Ie: newsfeeds.  I have f protocol enabled on my
: >mail link, and g on my news feed.

  If you can do it on a service-by-service basis, maybe a link-by-link
  basis?  Connection-by-connection would be best.  (I'd rather not
  worry about whether the remote admin. is likely to change modems.)

UUCP seems to work under the binary assumption of connect/not connect,
not connect with x attributes.  You COULD try separate connection attempts,
either through different modem register settings, or send/receive sequence
changes.  (It'd be nice if uucp had some kind of memory, and could adjust
itself based on system/line/method success, (so not to have so many
unsuccessful calls))

Any thoughts on this?

Kris A. Kugel    908-842-2707
hico2!kak    [e-mail]