SCRIBE DISTRIBUTION TAPE

This  tape  contains  the  Scribe document-formatting system level 2A(432) from
Carnegie-Mellon University.  If this tape is an object-only distribution,  then
it will contain five save sets; if it is a source distribution, it will contain
two more save sets for a total of 7.

Under  no  circumstances  is this data to be distributed by you to another site
without our explicit written permission.  Scribe is distributed without  charge
to  universities, but it is a commercial software product, and its unauthorized
distribution is liable to prosecution.  If this distribution  includes  source,
as  it  does  in  special  cases,  it  is  your  responsibility  to ensure that
unauthorized persons do not have access to the source:  it must be placed in  a
protected directory.

If you need to contact me, try (412) 578-3053, or write to


    Brian Reid
    Computer Science Department
    Carnegie-Mellon University
    Pittsburgh, PA  15213

The various save sets on this tape are allocated as follows:

   1. Save  set  INTRO:  the file Intro.doc, which is the file you are now
      reading.

   2. Save set OBJECT:  various versions of the program.

          SCRIBE.EXE      is the version we run here.
          SCR10.EXE       is a Tops-10 version with XGP driver omitted
          SCRB20.EXE      is a Tops-20 version (long filenames)
          SCR10.SHR,.LOW  is the same as SCR10.EXE, for older monitors

   3. Save set DBASE:  The data base.

   4. Save set EXAMPL:  Some sample manuscript  files  collected  at  CMU.
      These are included only as examples of Scribe usage.

   5. Save  set  RELOC:  Copies of the various .REL files that are used to
      link a running Scribe.  Also contains the linker command files.
If this is a source distribution, it will also contain:

   6. Save set SOURCE:  The source, which also contains a copy of Bliss-10
      level 7E.

   7. Save set DEBUG:  The experimental source.  This  save  set  contains
      the  entire  contents  of  my working directory in which I do Scribe
      development, and there will often be some random junk in there along
      with the program.

                               SCRIBE AT RUNTIME

The runtime Scribe system consists of a program, SCRIBE.EXE, and a  data  base.
This  data  base  is  about  a  hundred  ascii files; SCRIBE refers to them for
information about typographic style and format.  There are also a few "utility"
programs; PODTYP, for example, will print a Scribe-generated  .POD  file  on  a
Diablo (Use PODTYP.EXE on TOPS-20 and PODTYP.SAV on TOPS-10).

The file SCRIBE.EXE distributed on the tape is linked for a KI-10 or KL-10 with
memory  mapping.    It has the high segment starting at 650000, so it cannot be
used on a TOPS-20 or TENEX system, because the compatibility package  needs  to
be  at  that address.  The file SCR10.EXE is the same version of Scribe, but it
has been compiled without the XGP driver, which saves a little bit of core, and
it has been linked with the hiseg starting at 400000.  SCR10.SHR and  SCR10.LOW
are the same core image in the older two-file format.

The  file  SCRB20.EXE  is  a  semblance of a Tops-20 "native mode" version.  It
still has  a  Tops-10-type  command  scanner,  it  does  not  support  filename
recognition,  and  it  probably has a few rough edges in its treatment of files
and dates and things.  We run a Tops-10 system, and therefore don't  have  much
chance to debug the native-mode Tops-20 version of Scribe.

To get Scribe running at your site, you need to do three things:

   1. Allocate a PPN (Tops-10) or a directory (Tops-20) to hold the Scribe
      database.   It is not appropriate that [1,4] or <SYS> or <SUBSYS> or
      [1,2] or such account be used to hold the Scribe database.  Allocate
      a directory named <SCRIBE> or some PPN whose only purpose is to hold
      the Scribe database.

   2. Copy to this directory or PPN the entire contents of the third  save
      set on this tape (the save set whose name is DBASE).

   3. Retrieve  the various versions of the actual SCRIBE program from the
      second save set on the tape (save-set name is  SCRIBE),  rename  the
      appropriate  version  of  the program to be named SCRIBE, and modify
      the core image so that it knows to look  for  its  database  on  the
      database  account  that  you  have  just  set  up.    Then  put this
      newly-modified SCRIBE.EXE onto <SUBSYS> or [1,4].

We should probably elaborate on step 3, "Modify the core image...".  SCRIBE.EXE
contains reconfiguration code in it; all you have to do is tweak a variable and
then start the core image; it will slide into the reconfiguration code.

Get the core image with a GET command, and store the value 1 into location 140.
On TOPS-10 you do this by typing the commands

    E 140
    D 0 1

whereas on TOPS-20 you do this by typing

    D(escape)140(escape)1

Scribe will then take you through a dialog that looks something like this:

    *** Reconfigure SCRIBE.  Type strings, follow with CRLF ***

    1. Site Name.  Current string:
       "Carnegie-Mellon University"
       New site name:


    2. Site Code.  Current code "CMU"
       New site code:

    3. Library definition.  Type a template for finding database files.
       Use "#" for filename position, "@" for extension position.
       Current template is "DSK:#.@[X385BR10]"
       New template:

    4. Signon.  Current signon message is "Gripes to Reid@CMU-10A"
       New signon message:

    All done.  Be sure to save the core image.

In the file template, where we currently have [X385BR10],  you  would  type  an
octal  ppn.    X385BR10  is  a  silly  CMU  ppn;  it translates internally into
something like [55542,36234].

If Scribe is unable to find something in its database, it will let you know  in
no  uncertain terms.  Make sure that the files are sufficiently unprotected and
that the directory is sufficiently  unprotected.    Make  sure  that  the  disk
structure  name  in the file template corresponds to the disk structure name on
which you really put the database files.

                               THE SAMPLE FILES

The fifth save set on the tape (SSNAME EXAMPL) contains  a  number  of  example
files that were culled from here and there at CMU.  Most of them were developed
for  printing  devices that can print about 100 to 110 characters per line, and
they might possibly trigger "line too long" errors in places when compiled  for
export line printer devices.  Don't worry about it.

                     RECOMPILING AND RELINKING THE SYSTEM

If  this distribution includes the source, then the fifth save set on this tape
is named SOURCE.  It has the  various  .BLI  source  files,  the  various  .REQ
require  files,  and  a  couple of .MAC files for low-level support.  The BLISS
compiler, BLISS.EXE, is level 7E(224); if you have any level-7 BLISS  compiler,
you don't need this one.

Some versions of LINK complain a lot if there is a symbol declared EXTERNAL but
never  defined  or  referenced.   Ours doesn't.  Therefore, it may well be that
when you compile and link the system, you will  find  some  undefined  symbols.
Probably  not  to worry, but if SIX12 is one of them, you should edit the first
line of SCRCMU.BLI to remove the TIMER=EXTERNAL(SIX12) clause.

The file SCRIBE.CMD is the Tops-10 CCL command file that I use to generate  the
production version of the system.  Before you use it, you will probably want to
edit  the  file CONFIG.REQ to set the parameters in it to values that you like.
If you know how to use SIX12 and want the debug version of the  compiler,  then
SCRB.CMD  is  the command file to use.  As far as I know the TOPS-20 version of
CCL can't handle these files, so  if  you  are  running  on  TOPS-20  you  will
probably have to compile by hand instead of using command files.