ken@argus.UUCP (04/01/87)
In article <122@njitsc1.UUCP>, bc@njitsc1.UUCP (Bill Cheswick) writes: > In article <7319@boring.mcvax.cwi.nl> jack@boring.UUCP (Jack Jansen) writes: > >The point is, why should we use those icky uucp protocols that are > >hardly documented, and unused except in some proprietary sofware > >Because it is faster. Compare the transfer speeds of uucp and kermit: you only > get about half your baud rate in kermit file transfers. uucp does better. uucp may transfer data faster, but kermit has managed to jump the most formidable barrier I know of to date with ease: going from a Unix box to an IBM 4361 via a 7171 protocal converter. That is what I consider a very good trick. Furthermore Kermit runs on every major machine I've used, including some of the really odd combinations down at EIES. To my knowledge, uucp is limited to Unix and ms-dos. > Bill Cheswick uucp: bellcore!argus!njitsc1!bc > New Jersey Inst. of Tech. ARPA: argus!njitsc1!bc@flash.bellcore.com > bitnet: bc%argus.uucp at mouton.arpa" -- Kenneth Ng: Post office: NJIT - CCCC, Newark New Jersey 07102 uucp !ihnp4!allegra!bellcore!argus!ken *** WARNING: NOT ken@bellcore.uucp *** bitnet(prefered) ken@orion.bitnet Kirk: "I don't care if you hit the broadside of a barn" Spock: "Why should I aim at such an object?"
wmf@chinet.UUCP (04/02/87)
In article <122@njitsc1.UUCP>, bc@njitsc1.UUCP (Bill Cheswick) writes: >> In article <7319@boring.mcvax.cwi.nl> jack@boring.UUCP (Jack Jansen) writes: >> >The point is, why should we use those icky uucp protocols.... > >uucp may transfer data faster, but kermit has... >formidable barrier I know of to date with ease: [Unix to IBM4361] >a very good trick. Furthermore Kermit runs on [just about anything] ...and can be the basis of a "poor man's network" of a UNIX box and a bunch of PC's as terminals. The best of both worlds for low bucks, with equal (generally excellent) support for all versions. Because of the copyright held by Columbia University, the program is allowed to mature in a somewhat orderly fashion and behave consistantly. But your entire site can have this software for $10. (If you want the manuals.) -- +-----------------------------------------------------------------------------+ | Fortiter in re, suaviter in modo Bill Fischer | | ...ihnp4!chinet!wmf | +-----------------------------------------------------------------------------+
grodberg@kodak.UUCP (04/09/87)
While we are on the subject, I think it would be advantageous to include
protocols that allow for the corrent (and future) generations of
error-correcting psuedo-full-duplex high-speed modems. For example the MNP
class 6 modems include data compression, flow control, and error correction
all for the low price of being essentially half-duplex. This means that an
efficient protocol would use very simple error-detection/correction schemes
if any, not require and response from the receiver until the end of the
file transmission, and work in only one direction at a time.
I am not suggesting that any other communication modes be abandoned, but
am simply pointing out that current protocols such as kermit and uucp actually
are more of a liablility than a help when using these high-tech modems, and
since they are likely to be the wave of the future, it would be good to
add adequate support for them in addition to the others.
--
Jeremy Grodberg
Usenet: ...rochester!kodak!grodberg
Arpa: grodberg@kodak or kodak!grodberg@rochesterreggers@julian.UUCP (04/14/87)
I can't help but comment on this debate since we've been using Kermit for
well over a year as the transport protocol to link many different machines
together for the purpose of E-mail. Here's an abstract to a paper I wrote
about a year ago describing our system.
MLNET: Using Kermit to Build a Network
(Interim Solutions)
MLNET is an electronic mail network at The University of Western
Ontario which links a diverse set of machines (VM/CMS, VAX/VMS,
Tops-10, CDC/NOS, Unix, and Prime) and provides a gateway service
between several mail networks (BITNET, UUCP, CDN, MLNET, and others
through remote gateways). Kermit is a public domain protocol,
developed at Columbia, available for a diverse set of machines, and
typically used for PC to mainframe file transfer under user
control. The MLNET/MFS uses Kermit protocols, when available, to
transport mail messages -- machines with Kermit can be added to the
network with minimal programming effort. This paper reviews the
design of the MLNET/MFS and shows how Kermit solves a major problem
for our network. The system provides an attractive interim strategy
since it may be implemented on machines as small as the IBM AT
(under Xenix), connects over available asynchronous lines
(including modems), transports messages using available protocols,
and requires minimal programming effort to add new hosts.
--
Telephone: (519) 661 2151 x6026
CDN: {postmaster/reggers}@UWO.CDN
BITNET: {postmaster/reggers}@uwovax.BITNET
UUCP: {postmaster/reggers}@julian.UUCP (...!watmath!julian..)