Postscript font usage

[Mark Fardal <fardal(at)weka.astro.umass.edu>]

Hi,

I have a somewhat open-ended question.  I'm wondering: how many of you
use Postscript fonts in your direct graphics work?  What are your
experiences with it?


Advantages of Postscript fonts: aesthetics, and more importantly
legibility.  In so many talks, the speaker rambles on about a plot for
five minutes in front of a bemused squinting audience, until someone
finally gets the courage to ask, "excuse me, what are the axes?"
That tends not to happen with postscript fonts.

It is a little bit of a trick to use special symbols (like the 
"sun" symbol in astronomy).  You have to get the fonts, digest the files
for IDL, load them, access them properly, and then make sure they're
loaded into the output file so they'll print.  But it does work in the end.


Disadvantages of Postscript fonts:  lots of problems in the current
implementation.  That's not to say there aren't workarounds, but
it can be a pain.  (I am using version 5.2.1L, but I believe at least
some of these problems persist in later versions.)  For example :

The placement bug: putting symbols from user-defined fonts screws
up the calculation of string length.  This means your titles are
offset if you use special symbols.

The font-switching bug: it is not possible to use more than one
user-defined font at a time.  If you only need one special symbol
that's fine.  If you need more it can be trouble.

The clipping bug: postscript text is not clipped no matter what you
have set the clipping to.

The scientific notation bug: using Times-Roman at least, 
scientific notation uses a raised dot instead of an "x".  
Try plot, [0,1.e9] to see what I mean.  Looks rather silly.  
(My preferred workaround is never to use values
that require scientific notation in a linear plot.)


Anybody else have comments or problems?  RSI is planning to roll
version 5.5 out (sometime this summer?), and it would be nice if they
would get PS fonts in order before then.  It would probably help if we
could identify all the things they need to fix.

thanks,
Mark Fardal
UMass