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

# Re: Thoughts

• To: math-font-discuss@cogs.susx.ac.uk
• Subject: Re: Thoughts
• From: Michael John Downes <mjd@MATH.AMS.ORG>
• Date: Mon, 16 Feb 1998 09:17:15 -0500
• Reply-To: mjd@MATH.AMS.ORG
• Sender: mjd@MATH.AMS.ORG

Thierry Bouche:

> another thing, about radicals again. A friend of mine, typographer in
> the outer (xpress) world was very surprized by tex's behaviour to use
> PS rules for building the extensible part of the radical (instead of a
> glyph). I think his remark had to do with resolution dependant PS
> yielding an inconsistent result in a PDF viewed on screen. However, it
> seems obvious that a rule construct cannot be hinted, although a small
> bar could be: is there any good reason to have tex compute the rule
> geometry rather than just putting enough small overlapping bars as in
> some plain tex macro as \hrulefill (if memory serves) ?

You were probably thinking of \rightarrowfill (\hrulefill uses
\leaders\hrule).

I'm not exactly an expert in PS rendering, could someone explain how a
hinted bar is going to be rendered better than a lineto at low
resolutions? I don't see how there could be a significant difference
unless the PDF viewer is really brain-dead about rendering horizontal
and vertical lines. Doesn't it boil down in practice to choosing between
one-pixel thickness or two-pixel thickness, and shifting 0.5 pixel (or
less) in a vertical or horizontal direction to put the line on a pixel
boundary? Are you saying that Acrobat Reader *doesn't* do a good job of
this for plain old horizontal and vertical lines, but somehow can do
better with bar glyphs from a font?

By the way, I seem to vaguely recall that the main reason given by Knuth
for the strange height/depth of cmex radicals was that it ensured that
the base of the horizontal bar coincides with a pixel boundary in the
radical glyph.

Michael Downes <mjd@ams.org>