  
 				VAX BASED  			 CP/M 'MODEM7' COMPATIBLE 			  FILE TRANSFER PROGRAMS     F     This  document describes the file transfer system in use between aF     VAX 780 and an IBM PC.  This file transfer  system  is  compatibleF     with the CP/M MODEM7 file transfer protocols, and was developed inF     C using the public domain CMODEM program as the basis for the file     transfer portion.     F     Two  filter  programs are provided in the package to insert/deleteF     CR's in CR/LF combinations to match the VAX  conventions  on  textF     files.   These  routines  are  run  on  the  VAX,  and are invokedF     automatically by the use of the appropriate DCL procedures.  TheseF     DCL procedures should be used on text files only.  FTP  should  be1     used directly when transferring binary files.   '     The programs in use are as follows:    	$ FTP -x FileName  ? 	This routine handles file sending ('-s'), receiving ('-r'), or F         appending  ('-a')  from a remote system.  Direction of flow inF         the above commands is relative to the  VAX.   On  the  commandF         line  -x  is  one  of the above (-s,-r,-a) and FileName is the$         file to be sent or received.   	$ VAX2IBM InFile OutFile  	 ? 	Routine  to  convert LF-only line termination to the CRLF line C         termination required by CP/M and MS/DOS compatible systems.    	$ IBM2VAX InFile OutFile   ? 	Routine to convert CRLF line termination into the LF-only line F         termination required by the VAX RMS file system.  Also removes6         any CTRL-Z characters found (CP/M EOF marker).  -     The DCL procedures in use are as follows:    	$ FSEND FileName   ? 	Invokes  VAX2IBM to build a properly formatted text file, then 3         FTP to send that file to the remote system.    	$ FRECV FileName  	 ? 	Invokes  FTP  to  receive  a text file from the remote system, F         then invokes IBM2VAX to properly format that file for VAX RMS.  #     A couple of notes are in order:   F     The  FTP  routine is set up such that only ONE port is required inF     order to run it, the terminal that issues the FTP command  is  theF     same  one  that  sends/receives  the  file.   The  use of a simpleF     terminal emulator program to activate  the  VAX  is  all  that  isF     required.   The  protocols  are  set  up  for this form of delayedF     synchronization.  FTP has been tested at 9600 baud and should  notF     fall  behind.  When the VAX is heavily loaded, occasional timeoutsF     may occur.  These  timeouts  most  often  manifest  themselves  inF     checksum  or  other errors.  The timeouts are necessary for properF     synch of the protocol in case of errors, and may not be disturbed.  F     DEC  C  runtime will output the LF's to the file during the IBM toF     VAX conversion.  This differs  slightly  from  the  way  it  wouldF     appear  on  the VAX in RMS native form (null characters, with someF     form of char count as termination).  If file is then copied  usingF     the  DCL  COPY  command, the LF's are removed, and Null charactersF     and char counts are inserted in their place.  The  only  places  IF     have  found  where  the  conversion format will cause a problem is(     with the DIFF command, and with CMS.  F     With  DIFF,  use DIFF/IGNORE=SPACING to cause a comparison of fileF     contents only.  With CMS, edit the file  and  save  it.   The  new*     version will be in the correct format.  F     The  original  MODEM  program  was  developed  by Ward ChristensenF     several years ago, and has undergone  extensive  modifications  inF     the  user  interface.   The  protocol  has  not changed, except inF     spin-off versions that play around  with  CRC  instead  of  simpleF     checksums  for  verification.   These  are  normally identified as     MODEM9 and and above.     F     There  may  be  other (possibly better) protocols around, but thisF     one has become the de facto standard for the micro world.  If  youF     have  problems,  contact me before changing the protocol.  I couldF     actually have an implementation bug (shock!).  It has  been testedF     fairly extensively with at least two other  systems, but bugs  are     insidious.  F     Bugs,  enhancements  and  any  other wild ideas should be reported:     via mail to LOAFMAN (TCC01 or TCC02), or via phone to:           Kenneth W. Loafman5         work: (214)556-8889 (sometime during the day) 5         home: (214)245-2618 (before 10:00pm weekdays) 