W8SDZ@SIMTEL20.ARPA (Keith Petersen) (08/10/85)
Not being on the INFO-MODEM list, I am not privy to the discussion
that was mentioned on the message forwarded to INFO-CPM, however
I do have one question to ask:
Does this one byte change to the header information (to select
block size) make KMD incompatible with programs that currently
support the XMODEM protocol? Any changes made to a protocol
should be made in such a manner that the new editions may still
work with previous versions, that may not know of the new feature.
Could somebody please inform me about this? Does anyone think
that I have a valid point? If not, why? ...and, what about Naomi?
-Richard Hartman
max.hartman@ames-vmsb
The following message from CompuServe explains...
Date: Tuesday, 6 August 1985 18:16-MDT
From: CSTROM at SIMTEL20.ARPA
Re: Protocol wars continue
Paul Homchick thought there might be interest in this message:
#: 142754 S0/Communications
03-Aug-85 08:44:29
Sb: #142740-#Yam & r
Fm: Paul Homchick 71445,527
To: Pete Holsberg 70240,334 (X)
I would have thought that my stance on the wars was quite clear by
now, but at the risk of boring everyone to death, I'm going to state
it one more time. (Hit Cntrl-O NOW to skip this, if you wish!!).
At this point there is NO difference between the 1K implementations
except for Irv's additions of another 'receiving handshaking
character' the timing dependent nature of the initial handshake, and
the difference between "S" and "SK". This is in variance from the
YMODEM standard already implemented on MSDOS and UNIX systems.
If the CP/M community adopts the KMD/IMP protocol, I think it will
be very unfortunate for two reasons.
1) It adds complexity, outside of the checksum, to an already shakey
protocol which reduces its suitability for use over timesharing
systems and packet-switching networks. Also by adding ANOTHER
handshaking character, it continues the bad precedent of the C, and
invites further "improvements" via this extension mechanism, and
further degradation.
2) The existence of a split between the MSDOS / UNIX & CP/M
impementations helps to ensure that adoption of the 1K packet will not
be widespread, and could hasten the end of XMODEM as a low-end
standard by which everyone can transfer files. For reasons given in
(1), the M-U world is not going to adopt KMD. For reasons which I can
only characterize as well-meaning but shortsighted, sections of the
CP/M community have not adopted YMODEM. As you note, the commercial
protocols have not knocked each other out, but they HAVE kept any one
of them from becoming standard. CP/M users are now in the minority of
micro users, and it's going to get worse. I fear that KMD, with its
CP/M Ostrich outlook, will be counter productive to widespread
transfer of data via telecommunications.
To me, it is evident that these are significant reasons for taking a
strong stand. I hope it is clear that there is no hidden agenda here.