Returned network mail

Postmaster POSTMASTER%DICKINSN.BITNET at cornellc.cit.cornell.edu
Sat Oct 21 06:23:05 AEST 1989


This message was automatically generated.

Your mail message could not be delivered at Dickinson College in Carlisle,
Pennsylvania because the user address was not known at our site.

Addresses at Dickinson College are of the following format:

        username at DICKINSN       e.g., POSTMASTER at DICKINSN
                                  or  ALLAN_J at DICKINSN
                                  or  WOLTER at DICKINSN

Usernames are typically the first 8 letters of the person's last name,
frequently with the first initial added, and occasionally with an
underscore.  As a result, it would be difficult to guess someone's username.
If you do not know the address of someone at Dickinson, please send a
message to POSTMASTER at DICKINSN asking for help.

----------------------------------------------------------------------
The following diagnostic is the reason for the return:

        %BNET-W-NOSUCHRCVR, receiver LEYON_C cannot be located

Returned mail follows:
----------------------
Received: From PSUVM(MAILER) by DICKINSN with Jnet id 1830
          for LEYON_C at DICKINSN; Fri, 20 Oct 89 15:34 EDT
Received: by PSUVM (Mailer R2.03B) id 2520; Fri, 20 Oct 89 14:34:13 EDT
Date:         Fri, 20 Oct 89 13:32:15 CDT
Reply-To:     Unix-Wizards at BRL.ARPA
Sender:       Unix-Wizards Mailing List <UNIX-WIZ at NDSUVM1>
From:         "Charles F. Ashley" <LOGSAMS at A.ISI.EDU>
Subject:      IBM, HONEYWELL OCTAL CODES TO UNIX PROBLEM
X-To:         unix-wizards-request at sem.brl.mil
X-cc:         unix-wizards at BRL.MIL
To:           Chris Leyon <LEYON_C at DICKINSN.BITNET>

PROBLEM : UNISYS 5000/80 INTERFACING WITH IBM/HONEYWELL.

Here's a challenge to the UNIX professionals. My agency has hit a
programming snag in rewriting our programs from the HONEYWELL to
the Army's UNISYS 5000. On this side of the ocean, we are baffled
as to a solution:

Problem Definition:

THE PROBLEM OCCURS WHEN UPLOADING OR DOWNLOADING VARIABLE-LENGHT RECORD
FORMATS USING 9 TRACK MAGNETIC TAPES. THE IBM AND HONEYWELL USE A FOUR-BYTE
RECORD CONTROL WORD (RCW) THAT SUPPLIES THE LENGHT OF THE RECORD AND A
FOUR-BYTE BLOCK CONTROL WORD (BCW). IN SHORT, UNISYS READS BOTH BCW AND RCW
AS THEIR ACTUAL OCTAL REPRESENTATION CAUSING THE TAPE TO APPEAR AS ONE
CONTINUOUS HUGE RECORD WITHOUT INDIVIDUAL RECORDS.

Question: HOW CAN AN IBM OR HONEYWELL RECORD WITH OCTAL CODES BE READ INTO
THE UNISYS 5000 SYSTEM? LISTED BELOW ARE BOTH UNBLOCKED AND BLOCKED
PICTORIAL REPRESENTATIONS OF THE RECORDS FORMAT.

EXAMPLE 1:
UNBLOCKED RECORDS +-----------------------------------------+
                  | BCW | RCW | RECORD | BCW | RCW | RECORD |
                  |  4  |  4  |  300   |  4  |  4  |   200  |
                  | 308 | 304 |record 1| 208 | 204 |record 2|
                  +-----------------------------------------+
THE RCW CONTAINS THE LENGHT OF THE RECORD PLUS ITS OWN LENGHT OF 4.
THE BCW CONTAIN THE LENGHT OF THE RCW PLUS ITS OWN LENGHT OF 4.

EXAMPLE 2:
BLOCKED RECORDS   +--------------------------------------------------+
                  | BCW | RCW | RECORD | RCW | RECORD | RCW | RECORD |
                  |  4  |  4  |  300   |  4  |  200   |  4  |  250   |
                  | 766 | 304 |record 1| 204 |record 2| 254 |record 3|
                  +--------------------------------------------------+

THE RCW CONTAINS THE LENGHT OF THE RECORD PLUS ITS OWN LENGHT OF 4. THE
BCW CONTAINS THE TOTAL LENGHT OF THE RCWS AND RECORDS PLUS ITS OWN
LENGHT OF 4.

THESE 4-BYTE WORDS (BCW OR RCW) ARE USUALLY IN THE OCTAL FORM. OUR COBOL
COMPILER (FAST/PHILON) DOES NOT SUPPORT 9 TRACK (OR IS NOT DOCUMENTED) AND I
HAVE FOUND NO OTHER SYSTEM UTILITY TO DUPLICATE THIS. I NEED YOUR
HELP / ASSISTANT / GUIDANCE.

If possible send replies both to the bboard and logsams. Time for us is short.

Charley F. ?Ashley, US ARMY
-------



More information about the Comp.unix.wizards mailing list