From ilh@goldilocks.LCS.MIT.EDU Tue Jun  9 03:38:45 1992
Received: from winnie-the-pooh.LCS.MIT.EDU by goldilocks.LCS.MIT.EDU (4.1/SLS-4.1.1)
	id AA08081; Mon, 8 Jun 92 15:38:30 EDT
Date: Mon, 8 Jun 92 15:38:30 EDT
From: ilh@goldilocks.LCS.MIT.EDU (Lee Hetherington)
Message-Id: <9206081938.AA08081@goldilocks.LCS.MIT.EDU>
Received: by winnie-the-pooh.LCS.MIT.EDU (4.1/SMI-4.1)
	id AA00827; Mon, 8 Jun 92 15:38:29 EDT
To: Law Hon Man <hmlaw@cucs18.cs.cuhk.hk>
In-Reply-To: your message of 9 Jun 92
Subject: Notes
Reply-To: ilh@lcs.mit.edu
Status: R

TechPhonetic was designed by Rob Kassel at the MIT Spoken Language
Systems Group.  It is based on GoudyOldStyle from Altsys.  It is
freely distributable but not public domain.

Kerning
-------

If you're using the kerning (specified in the AFM file) you may be
surprised.  There is a substantial amount of kerning that falls in the
following classes:

1) Superscripts for on- and off-glides are designed to kern with
   adjoining characters.  If you want to disable this kerning insert a
   non-breaking space character (202).  A singlenon-breaking space
   adds no horizontal space per se.

2) Diacritics meant to be centered over/under symbols will
   automatically center via kerning.  Again, non-breaking space
   characters can be inserted to disable this.  After such centering,
   you'll need to insert enough non-breaking spaces to space over to
   where the main character left off.  You'll have to look at the
   width differences (always a multiple of ten milli-ems) and divide
   by the width of the non-breaking space (five milli-ems) and further
   divide by two.  In addition, diacritics will center with other
   diacritics.

This is very preliminary documentation.

Lee Hetherington
Massachusetts Institute of Technology
Laboratory for Computer Science
Spoken Language Systems
ilh@lcs.mit.edu



