cal@JHUNIX.BITNET (cal) (10/07/89)
We use troff (ditroff) and Adobe's Transcript package for most of our
printing. I have written or modified a number of Postscript fonts for
non-European languages and have gotten them up and running in ditroff.
The only problem that I have yet to figure out a solution to arises,
apparently, in how ditroff does its metrics work.
Due to the number of characters in the fonts I have developed, som
of the characters have had to be encoded in high ascii position and
called with the ditroff string \(xx (xx replaced with the 2 character
code for the desired character). Each character in the font has its
own metrics value in the Postscript prologue file and in the
corresponding afm (Adobe font metrics) file. This latter file is
converted to a ditroff font description file with the afmdit program
supplied with Transcript. All of these files have the correct metrics
information for all characters in the fonts. The problem comes with
ditroff takes the original troff coded file and produces the
intermediary ditroff file. In that file, the high ascii (and only the
high ascii) characters of my fonts are not given the metrics value
from the ditroff font description file for that font, but the metrics
of the character of the same ascii position in the Times-Roman font
(the default font for ditroff). The low ascii characters are spaced
correctly, so ditroff is finding the font description file, and all of
the characters, low and high ascii, are printing correctly, with the
exception of the messed up spacing of the high ascii characters. If
anyone has an explanation, or better a solution to this problem, I
would greatly appreciate knowing about it.
Appended here are some samples of the various files illustrating
the situation in case my explanation above is not clear enough. For
anyone already familiar with these files, I apologize for this.
However, if there is something wrong in these files that I have
overlooked, seeing the pieces of code below might prompt a solution.
Steve Bennett
cal@jhunix
Postscript (from Syriac prologue file)
Encoding % ditroff octal
dup 66 /bP put % B
dup 77 /mP put % M
dup 162 /bF put % \(ct 242
dup 174 /TB put % \(fi 256
Metrics
/bP 912 def
/mP 816 def
/bF 614 def
/TB 896 def
from Syriac AFM file
C 66 ; WX 912 ; N B ; B 0 -24 912 360 ;
C 77 ; WX 816 ; N M ; B 0 -200 816 448 ;
C 162 ; WX 614 ; N cent ; B 0 0 564 360 ;
c 174 ; WX 896 ; N fi ; B 0 -336 896 768 ;
Ditroff description file Syriac font
(the metric value for each character -- 2nd column -- is derived by
dividing the Postscript value by 5)
B 182 2 0102 B
M 163 2 0115 M
ct 123 2 0242 cent
fi 179 2 0256 fi
Ditroff values in Times-Roman font description file for same characters
B 133 2 0102 B
M 178 2 0115 M
ct 100 0 0242 cent
fi 111 2 0256 fi
Section of troff file and ditroff file (done at 25 point so that the
character size matches the unitwidth of our printer.)
(troff)
.fp 4 sy \" sy is name of downloaded font
\f1 B\(ct\(fiMB
\f4 B\(ct\(fiMB
(ditroff)
x font 4 sy
f1 Printing Times Roman line
H555
wV792
50Bh133Cct
h100Cfi
h111cM
h178cB
n216 0
H555
wf4 Printing Syriac line
V1008
50Bh186Cct 186 is correct for Syriac "B", 133 in Times
h100Cfi Should be 123, not 100
h111cM Should be 179, not 111
h163cB 163 correct for Syriac "M", 178 in Times
n216 0
x trailer
V6336
x stopmackay@MANTA.CS.WASHINGTON.EDU (Pierre MacKay) (10/13/89)
If you are using Transcript, I am a bit puzzled to learn why you use ditroff instead of ptroff. Ptroff goes right back to the old C/A/T 432 units/inch resolution, and lets the Postscript RIP do the scaling. Simple and effective. Not the solution I would use for TeX, but it works fine for all my Troff (manual page) needs. I have one quarrel with ptroff. The prologue that it adds to ps output uses the value 432 (exactly the old C/A/T resolution) to set dimensions and you can see the results in -man format where the bottom line goes off the paper. The problem is evidently rounding in the RIP. Change the 432 to 440, (it's a divisor) and the output shrinks to just about the correct size. This is true for both the older and the newer LaserWriters. You may have to experiment with other PostScript processors. Email: mackay@cs.washington.edu Pierre A. MacKay Smail: Northwest Computing Support Center TUG Site Coordinator for Thompson Hall, Mail Stop DR05 Unix-flavored TeX University of Washington Seattle, WA 98195 (206) 543-6259
rcd@ico.isc.com (Dick Dunn) (10/17/89)
mackay@MANTA.CS.WASHINGTON.EDU (Pierre MacKay) writes: > If you are using Transcript, I am a bit puzzled to learn why you use ditroff > instead of ptroff. There are good reasons for using ditroff (and psroff) rather than old troff and ptroff (the C/A/T interface). Ditroff works with grap. It works better than old troff with pic. Font support for multiple families actually works well. You can have more fonts "mounted" (available) at one time. Point size selection is better. Ditroff is also a lot cleaner internally, for whatever that's worth. -- Dick Dunn rcd@ico.isc.com uucp: {ncar,nbires}!ico!rcd (303)449-2870 ...No DOS. UNIX.