M2 KERMIT is  a modification of KERMIT  UCSD (UCMICRO version
    by  Tim Shimeall  of the  University of  California, Irvine ,
    which in turn is a  modification of UCTERAK by Kate MacGregor
    of Cornell University). M2 KERMIT  was created by Henry Balen
    of  Lancaster University,  Lancaster to  run on  a Magiscan 2
    image  processing  system (made  by  Joyce  Loebl). Like  the
    previous version this is written entirely in UCSD Pascal, and
    minor changes were needed in  the translation from UCMICRO to
    M2 KERMIT.  This version  does have  a few  machine dependant
    features but these are in seperate units.

    The following files make up M2 KERMIT;

       BINUTILS.TEXT
       DISK.TEXT
       FILEUNIT.TEXT
       HELP.TEXT
       KERMIT.TEXT
       PARUNIT.TEXT
       RECSW.TEXT
       RS232.TEXT
       RSUTILS.TEXT
       SENDSW.TEXT
       SYSUNIT.TEXT
       UTILS.TEXT
       HANDLE.TXT
    All these files have UCJ  prefix on front in the distribution
    directory, with the first 3 characters of their names following.
    The    files    DISK.TEXT,    FILEUNIT.TEXT,    PARUNIT.TEXT,
    SYSUNIT.TEXT are separately compiled  units (modules) for use
    by the main program KERMIT.TEXT;  the rest are the text files
    that make up M2KERMIT.

    The  Magiscan  2   does  not  have  the   capability  to  use
    interrupts, so this  version is like the  UCMICRO version and
    makes  use of  polling. Unlike  the microengine  the Magiscan
    does  have  enough memory  so  the  command parser  has  been
    'reconverted'  from   a  "segment  procedure"  to   a  "unit"
    (PARUNIT.TEXT). So far the time  out mechanism is the same as
    for the  microengine; I  hope to make  use of  the Magiscan's
    internal clock in  a future version. There are also  a few of
    other units  which are machine dependant:  The RS232, SYSUNIT
    and  the  DISK  unit;  the latter  was  written  because  the
    Magiscan's  Image load  and  save routines  didn't cope  with
    eight bit images, that was solved in this unit. Though a word
    of warning has to inserted here, DO NOT KRUNCH AN IMAGE FILE,
    the UCSD filer on the magiscan seems to upset the way I store
    the images i.e. they lose a bit.

    M2 KERMIT has the following limitations:

    a) No wild card designations of file names
    b) No character repeat counts
    c) No  '?' and <esc> at the end of a command line.
    d) No server communications
    e) Only one  file is sent at  a time (a break  packet is sent
    after each file).




    KERMIT M2 has been thoroughly  tested in connections with VMS
    and BBC KERMIT, but errors  may persist. In particular, since
    there  are  no interrupts  in  this  version, characters  may
    occasionally  be  lost.  Character  loss  was  observed  with
    transfers from the BBC and the MUX  delay had to be set to >=
    4; this enabled a wait between transmision of characters from
    the BBC. A  way round this would be to  implement the receive
    and send  packet routines in  machine code, and  possibly the
    terminal emulator, this would speed  up the polling and hence
    reduce the likelyhood of character loss.

    The commands recognized by KERMIT  M2 are a superset of those
    for  KERMIT UCMICRO.  I shall  give the  descriptions of  the
    commands that  I have added and  you should refer  to UCMICRO
    and TPKERM for the rest:

     SET BAUD  <rate>: this command enables  the user to  set the
    Baud rate,  the value  of <rate> can  be examined  by doing a
    SHOW BAUD (or a SHOW ALL).

     SET DISK  <vol>: this  command enables the  user to  set the
    work disk to any valid  UCSD block device (i.e #4,#5,#9,#10),
    the value  of <vol> should be  any of the  following 4,5,9,10
    dependant on  which devices your machine  has available. This
    can also be examined by use of the SHOW command.

     DIRECTORY: this  command gives the directory  of the current
    work disk, it gives an extended directory listing.

     DELETE <filename>: this command deletes the given <filename>
    from the work disk.

     TRANSFER TYPE  <type>: this tells  KERMIT what type  of file
    transfer is going to take place, the choice of transfer types
    are TEXT,  DATA, CODE, IMAGE. The  first is the  default, for
    IMAGE types an image has to be loaded using the LOAD command.

     LOAD  <filename>: this  loads an  image (called  <filename>)
    from the  work disk, this  can only  be done if  the transfer
    type is IMAGE.