[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]


------- Forwarded Message

Received: by goofy.zdv.uni-mainz.de with SMTP ; Thu, 5 Aug 1993 15:54:11 +0200
Received: from MATH.AMS.ORG by MATH.AMS.ORG (PMDF #2306 ) id
 <01H1DQLMWYBKLQV6XS@MATH.AMS.ORG>; Thu, 5 Aug 1993 09:49:30 EST
Date: 05 Aug 1993 09:49:29 -0400 (EDT)
>From: Michael Downes <MJD@MATH.AMS.ORG>
Subject: Re: accent widths [note change of Subject]
In-Reply-To: <9308031137.AA00822@micky.zdv.Uni-Mainz.DE>
To: ziegler@micky.zdv.Uni-Mainz.DE
Message-Id: <744558569.944045.MJD@math.ams.org>
Content-Transfer-Encoding: 7BIT
Mail-System-Version: <MultiNet-MM(348)+TOPSLIB(156)+PMDF(4.1)@math.ams.org>

 You wrote :

 For example: it would be nice to have ten or
 twelve different widths of widehat and widetilde symbols, not just four
 or five.

  8 different sizes for all sizeable accents. Is that not

Maybe. How wide will the widest accent be? Enough for four letters
(average width) ? five letters?  Suppose a math italic font has
characters and widths as follows:

  i 3pt
  n 4pt
  m 5pt
  R 7pt
  M 10pt

It's clear that various combinations of these letters will cover a
wide range of widths.  How accurately must the width of the accent
match the width of the covered letters, to achieve adequate quality?
Or high quality? That is, is one point of shortfall or overlap on
either side acceptable? Is two points of shortfall/overlap acceptable?
Does the amount of acceptable shortfall/overlap vary according to the
shapes of the covered letters? I don't think I am particularly well
qualified to judge these questions, but I would like to see a little
documentation describing these decisions included in the documentation
for the new math font encoding.

------- End of Forwarded Message

any comments ?