Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!utgpu!water!watnot!watmath!clyde!rutgers!ames!ucbcad!ucbvax!decvax!decwrl!sun!imagen!atari!apratt
From: apratt@atari.UUCP
Newsgroups: comp.os.minix,comp.mail.uucp
Subject: Re: uucp source copyright status - IMPORTANT
Message-ID: <692@atari.UUCP>
Date: Thu, 2-Apr-87 15:59:36 EST
Article-I.D.: atari.692
Posted: Thu Apr  2 15:59:36 1987
Date-Received: Sun, 5-Apr-87 03:07:33 EST
References: <1643@ttrdc.UUCP>
Distribution: comp.os.minix
Organization: Atari Corp., Sunnyvale CA
Lines: 22
Xref: utgpu comp.os.minix:488 comp.mail.uucp:400

in article <1643@ttrdc.UUCP>, levy@ttrdc.UUCP (Daniel R. Levy) says:
> 
> Can Kermit work auto-dialers and negotiate the login protocol?  Also--does
> Kermit negotiate bad-data retry on a smaller scale than whole files?  Just
> curious (I'm not a Kermit guru as you can plainly see).

Kermit would not be called upon to negotiate logins: what people are
talking about is placing Kermit at the file-transfer level rather
than uucp.  The login protocol layer would be the same in both cases.

Kermit handles bad transmissions at the packet level; unfortunately,
packets are only 94 data bytes long, with (I think) about 6 bytes
overhead.  Finally, Kermit is a window-of-one protocol, meaning that
each packet must be ACKed or NAKed before the next one starts.  The
overhead of having short packets and many, long ACKs is bad enough, but
with some 2400 and 9600-baud modems turnaround (from send mode to
receive mode) takes as much as 1/2 second.  Definitely bad news.

/----------------------------------------------\
| Opinions expressed above do not necessarily  |  -- Allan Pratt, Atari Corp.
| reflect those of Atari Corp. or anyone else. |     ...lll-lcc!atari!apratt
\----------------------------------------------/