HISTORY of MISH:

    The MASH program:

	The  MISH  program  was  created from a program called MASH.  The
    MASH  program  was  used  to convert and create a file which could be
    transmitted  on a HASP reader stream.  These files had to be 80 bytes
    bytes  in length and could contain binary data.  The MASH program has
    the following features:

	o  Uses HASP data compression.
	o  Allows conversion of any type of file.
	o  File attributes stored in a header record.
	o  Format acceptable for HASP and 2780 punch streams.
	o  Written in RSX MACRO-11.
	o  Runs on RSX-11M/M+ and VMS in compatibility mode.


    The MISH Program:

	There was a need to transfer files on the HASP reader streams and
    also  to  be able to transfer these files using the RSXNET and VAXNET
    programs.  It  seemed  logical then to convert the output of the MASH
    program  into printable ASCII to accomplish this need.  At this time,
    the  MASH  program was converted into MISH.  The MISH program has the
    following features:

	o  Uses HASP data compression.
	o  Allows conversion of any type of file.
	o  File attributes stored in two header records.
	o  Format acceptable for HASP and 2780 reader streams.
	o  Format acceptable to the RSXNET and VAXNET programs.
	o  Format acceptable to the VMS Code Management System.
	o  Base character for the ASCII conversion is '%' ('$'+1)
	   to allow easy file inclusion in VMS batch jobs.
	o  Conversion is 2 bytes into 3 printable ASCII bytes with
	   a maximum record size of 78 bytes (reader stream max=80).
	o  Written in RSX MACRO-11.
	o  Runs on RSX-11M/M+ and VMS in compatibility mode.


    VMS MISH Program:

	All  seemed  well  with the MISH and MASH programs until VMS V4.0
    was released.  Now the Application Migration Executive (AME) used for
    running compatibility programs on VMS became a layered product called
    VAX-11/RSX.  Existing  sites  could  obtain the first release of this
    product free of charge if they were upgrading from a previous version
    of VMS, otherwise there was a fee.  There was also a need to run MISH
    on the Micro-VAX which doesn't include compatiblity mode.

	Thus,  MISH needed to be rewritten to run on any VMS system.  The
    language  chosen was the new FORTRAN available on VMS V4 since it now
    contains  extensions which allow easy access to the RMS structures to
    obtain  the file attributes.  Also, the source modules created can be
    easily included in the VAXNET program also written in FORTRAN.
    
	Since  the MISH and MASH programs had much common code, I decided
    to  include the features of both programs in the VMS MISH.  To select
    MISH  mode,  which is the default, the /ASCII qualifer is used and to
    select  MASH  mode, the /BINARY qualifier is used.  I also discovered
    during  the  rewrite that the compression routine could be optimized.
    This optimization resulted in the following savings when used with an
    executable file during testing:

    OLD:
	Program  Blocks In  Blocks Out  Records Out
	-------  ---------  ----------  -----------
	MISH        30          42         283
	MASH        30          29         174

    NEW:
	Mode     Blocks In  Blocks Out  Records Out  Savings
	----     ---------  ----------  -----------  -------
	ASCII       30          40         264         19
	BINARY      30          27         168          6

     Since most file transfers are done at slow speeds (300 or 1200 baud),
     the above savings will be significate on larger files.


     XMODEM Qualifier:

	It's  obvious from the above statistics that conversion of binary
     data  to  ASCII is expensive.  Since the VAXNET program supports the
     XMODEM  protocol, I decided to allow the option of generating a file
     suitable  for transfer using XMODEM in binary mode.  When the XMODEM
     qualifier  is specified, BINARY mode is selected and the record size
     is  changed to 128 (size of an XMODEM data block).  Since the XMODEM
     protocol  has  better  error  checking  than the VAXNET protocol and
     allows binary transfers, I prefer to use this program.


     Design Notes:

	o  The  header  record  is not compressed which forces one or two
	   header  records  for  MASH  and MISH respectively.  The header
	   record should be compressed just like any other record.

	o  The UIC stored in the VMS header record is only a word instead
	   of a longword to be compatible with the PDP-11 programs.  This
	   UIC is used for identification of where the file came from, it
	   is not used when decompressing the file.

	o  The  maximum size of the file name stored in the header record
	   is limited to 31 characters.  This limitation is imposed to be
	   compatible  with  the  PDP-11  programs.   This causes no real
	   problem  except  that the file name in the header record can't
	   be used when the file is decompressed. In this case, an output
	   file name must be specified. During compression, anything over
	   the  limit  is  overwritten  with the file attributes which is
	   much more important than the file name.

	o  Access Control Lists (ACL's) are not retained.