Wait, does more informality mean that individual glyphs for the same character can be different even within the same sentence?

Unfortunately not, that would require a random axis, or a contextual swapping based on adjacent letterforms.

Prof. Hermann Zapf's eponymous Zapfino has the latter --- I even included an animation of it in my paper on it:

http://ftp.tug.org/TUGboat/tb24-2/tb77adams.pdf

Whoa, opulent!

Thanks!

It was a fun paper to write, but came a bit too late to have any influence --- at the same conference Jonathan Kew presented XeTeX and shortly thereafter luatex was developed, so it ceased to be necessary to stitch together hundreds of .eps files to make a possibly several GB PostScript file which then had to be distilled to a PDF using the commercial Adobe Acrobat.