[texhax] How can we best keep on using small caps and oldstyle figures?

[Pierre MacKay]

Paul,

Your detailed response was (in one sense) exactly the sort of thing I 
was hoping for, but I am distressed, as obviously you are, by some of 
your observations.

Immediate relief for my immediate situation came when the online 
distiller opened up again, and the Windows instance of Distiller was 
silently corrected (which means that Adobe is secretly invading my 
system to fix things, and that does not entirely sit well with me).

Why do I use distiller? Because the commercial firms I have to satisfy 
won't accept anything else.  Moreover, not all PDF is the same.  For 
example, Ghostscript produces apparently successful PDF through the 
various ps2pdf instances, but you cannot display them in Adobe Reader 
7+, because Reader 7+ substitutes blank space for all the characters in 
Adobe Expert Encoding.   The various direct to PDF versions of TeX may 
help, and ultimately I may have to try them, but the manuals are not 
reassuring.  The correspondence associated with direct-to-PDF output 
does not sound as if refinements such as small caps and old -style 
figures are of any more concern to the participants than they are to the 
"industry" as you describe it.

In addition, I have looked through the documentation to find any support 
for the independent setting of bleedbox and trimbox that seems to be 
offered only by a computer resident version of distiller (not offered by 
online distiller) and when one makes up a cover with a solid color 
underlay, printers insist on a bleedbox 18pt outside the trim box all 
round to insure against white hairlines at the edge of the cover. I 
could force this code into a PDF that didn't have it through a perl 
script, but it would be rather a bear to do, and Adobe Reader does have 
the virtue that it shows you that trimbox is entirely inside bleedbox 
with a discrete green line that is not part of the PDF.

Free font development would be nice but, as you say, it is rarely done 
to anything like the standards that some of the older Type1 packages 
maintained.  One of the problems that shows up again and again is 
completely random set-widths.  In one once-touted Greek package I found 
/alpha, /alphaacute, /alphagrave, and /alphaperispomene each with 
different widths, the most extreme pairs reaching nearly a point in 
difference.  That really shows up in densely set text.  (It took me 
nearly four years to refine Ibycus to even its present imperfect level.)

I recently had the pleasure of setting a long text in Bruce Rogers's 
Centaur, and I doubt that any free clone will ever come near the 
elegance Rogers put into that.  It may be the finest single font ever 
produced in the United States. Centaur is not a good screen display 
font, but it seems grossly wasteful of quality to dump it forever just 
for that limitation.

Perhaps, as you suggest, luatex may provide a way of holding on to the 
good fonts from the past.  When I get out from under my present backlog, 
I shall look into it, and be very grateful to you and luatex for the 
assstance.

Best regards,

Pierre