Path: utzoo!utgpu!jarvis.csri.toronto.edu!clyde.concordia.ca!uunet!cs.utexas.edu!sun-barr!decwrl!shlump.nac.dec.com!minim.dec.com!batcheldern
From: batcheldern@hannah.enet.dec.com (Ned Batchelder)
Newsgroups: comp.lang.postscript
Subject: Re: PostScript utility for font changes, subscripts within string
Message-ID: <7026@shlump.nac.dec.com>
Date: 20 Dec 89 16:09:04 GMT
References: <12294@phoenix.Princeton.EDU>
Sender: newsdaemon@shlump.nac.dec.com
Distribution: na
Lines: 51

Rather than limit yourself to the idea that your data representation must be
a string, which is inert and must be decoded, pick a more convenient
representation for your text-with-font-changes. I've done something like
this and it was very easy.

Text is represented as a procedure which when executed shows the text. For 
example:

	/tr /Times-Roman findfont def		% Base font
	/rm tr 10 scalefont def			% font for Roman
	/sc tr [ 8 0 0 8.5 0 0 ] makefont def	% font for small caps

	{ rm setfont (P) show
	  sc setfont (OST) show
	  rm setfont (S) show
	  sc setfont (CRIPT) show }

The procedure at the end is the representation of the caps/small-caps form
of the word "PostScript". Now to show this text, you define "textshow":

	/textshow /exec load def

To compute the width of the text, you have to be a little trickier. Nulldevice
comes in very handy here:

	/textwidth {
		gsave nulldevice
		0 0 moveto
		textshow
		currentpoint
		grestore
	} def

Then you can center text:

	/textcentershow {
		dup textwidth pop
		-2 div 0 rmoveto
		textshow
	} def

Etc.

It isn't a very compact representation, but it sounds like your application 
won't be doing much text anyway, and it can be made compacter by the proper
definition of procedure to use within the text itself.

[I typed this in off the top of my head, so there may be errors, but the ideas
are sound: I've used it and it works.]

Ned Batchelder, Digital Equipment Corp., BatchelderN@Hannah.enet.DEC.com