Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!husc6!hao!boulder!fesk!sverre
From: sverre@fesk.UUCP (Sverre Froyen)
Newsgroups: comp.text
Subject: Re: Bad spacing in old troff
Message-ID: <158@fesk.UUCP>
Date: Sat, 14-Nov-87 15:40:13 EST
Article-I.D.: fesk.158
Posted: Sat Nov 14 15:40:13 1987
Date-Received: Sun, 15-Nov-87 20:28:15 EST
References: <3659@trwrb.UUCP>
Organization: SERI, Golden, CO
Lines: 28

in article <3659@trwrb.UUCP>, simpson@trwrb.UUCP (Scott Simpson) says:
> My version of old troff doesn't always space large mathematical symbols
> correctly when using eqn and I was wondering if anybody had a fix.  For
> example, if I print the document ``Typesetting Mathematics - User's Guide''
> using old troff, the Sigma in the equation on the first line of the abstract
> is shifted over to the right about the same distance it is wide.  It looks
> something like
> 		.   S
> 		k>=1
> 
> where S represents the Sigma and the period is the reference point where it
> should be.  Ditroff does not suffer from this malady.

The CAT typesetter requires a shift of 55 units when inserting or
removing the so-called doubler lens (used to create large point sizes).
The output of troff contains this shift which has to be removed by
the post-processor. The sum sign is a large SIGMA (~point size 15),
involves the doubler lens, and therefore the shift. My guess would
be that the incorrect position is caused by your your post-processor
correcting for the shift in the wrong direction (or not at all).
Here I speak from experience having done this incorrectly myself
in an early version of the LaserJet driver recently posted to
comp.sources.misc.
-- 
Sverre Froyen
UUCP:   boulder!fesk!sverre, sunpeaks!seri!fesk!sverre
ARPA:   froyen@nmfecc.arpa
BITNET: froyen@csugold.bitnet