October 28, 1985

To:  Bill Saphir
     David Grubbs
     Pete Kelley
     Roger Roach
     Leo Ryan
     Alix Vasilatos

From: Linda B. Merims, Athena Scribe Database Administrator

Subject:  Project Athena/SIPB/IS  Xerox 9700 Service


We are putting an experimental, unsupported version of Xerox 9700 printer
support for both line printer and Scribe output out immediately in
the /usr/unsupported/sipb directory.  We intend that this software
and service will become supported in January, 1986.  Since this is
a significant service that depends upon three different administrative
bodies to work, I outline here what the pieces of this chain of
service are, whose responsibility they are, and what I believe
each are supposed to work like.  

Please comment.


Purpose

We want to provide Xerox 9700 printing service for these reasons:

 - A high-volume printer to backup the flagging LN01's
 - A cost-recovery printer
 - A "typesetter-like" variable width font printer
 - A replacement for the SIPB-funded 9700 thesis output service
   discontinued this fall

The 9700 will not answer the need for graphics output.  It can't do
it now, and @i(in principle) won't ever be able to.

We have no current plans to force use of the 9700 (for example, for
print jobs longer than 20 pages).


Overview

Xerox 9700 service is provided this way:






I now take each numbered block and describe its function, components,
responsible party, and future direction.


1. Plain Text Output

Any user can send ascii text files to the Xerox 9700.  It will behave
like any line printer.  The user will have options.  These are
described below under "Modified lpr Command."


2. Scribe Output

To make Scribe produce X9700 output, there are several components.

@subsection(Xerox 9700 Driver License)

IS has purchased a campus-wide X9700 driver license from Unilogic.
To keep working, Scribe must have an up-to-date @i(/usr/lib/scribe.sit)
file that contains the proper codes to enable the 9700 driver.
"Site files" periodically expire.  Unilogic sends a letter listing
the new site codes to IS in the person of RK Scott, who in turn forwards them
to the other MIT sites.  {Who gets them at Athena?  Me?}
Athena operations under Alix Vasilatos should install the
new site files on Athena machines.  Not everybody at MIT has the same repatoire
of licenses (for example, Athena also has an Imagen license), and hence
somewhat different site codes.
If a site file has expired altogether, no user can run Scribe.  If
the codes for drivers are incorrect, the individual driver will not
work.


@subsection(Scribe Database)


Scribe uses a large directory of files to interpretively determine
how to build output documents.  This is the "Scribe Database."  The
database is organized into layers of definitions, from the abstract
such as "a report" to the very concrete such as "an italic in
helvetica 10 is the typecase h110jpIt
is the directory @i(/usr/athena/lib/scribe/database).
You get all the files necessary to make a 9700 work when you
get the basic Scribe license.  (It just won't work until you get
the correct site codes when you purchase the 9700 license.)

The files involved with making the 9700 work are:

Every document type has a @i(.mak) file.  (E.g., report.mak, articl.mak, etc.)
It describes the overall way that document type should look, such
as whether or not it has a title page.  .mak
files are subdivided with @@marker statements, one for each type
of output device.  These subsections describe how the document type
should look @i(on each particular device).  

No changes have been make to the @i(.mak) files from Unilogic.


Each device has a @i(.dev) file.  The device file for the 9700 is
called x9700.dev.  Device files define what each Scribe environment
(text, format, example, description, heading, etc.) looks like
(width, indent, font) on that device.  It also defines the file type
the output file will be named with.  Bill Saphir made one or
two minor changes to the x9700.dev file.  I don't know what they were.


The next layer down is the .fon files.  finger