The Civil Cyrillic Alphabet, introduced by Peter the Great, makes many Russian letters look like some Latin ones; this is a nuisance for the programmers: what looks identical on the screen can be quite different in the code; moreover, some letters are identical in one case (T,Т y,у) and different in the other (t,т Y,У). The Western programmers know a similar problem with O/0 and 1/l/I disambiguation.
In a Unicode font the problem is even worse, because the Latin/Cyrillic ambiguity is complicated with the Greek. Still, I've tried to modify a couple of the bitmapped BDF Unicode fonts by Markus Kuhn which I use on Sparc computers so that I can see the difference and guess which alphabet a given letter belongs to. Actually these are 9x18 fonts in the plain and bold flavors:
-Pok-Fixed-Medium-R-Normal--18-120-100-100-C-90-ISO10646-1
-Pok-Fixed-Medium-R-Bold--18-120-100-100-C-90-ISO10646-1
Markus Kuhn, the initiator of the project, dislikes this idea. He wrote in his advice to contributors:
Always look first at existing similar characters in order to preserve a consistent look and feel for the entire font and within the font family.
In principle this is reasonable, provided that “consistent” does not
imply “identical”. The Gill Sans font is extremely consistent in
that it identifies 1Il as well as Oo; here is a specimen
from Wikipedia:
But this is unacceptable in a programmer font, and the fixed font
family in question makes the corresponding glyphs distinguishable.
Similarly the English “a copy” should be distinguishable from the
Russian “а сору” (e.g. «а сору было много») — such words are often
confused in automatic text recognition and require manual correction).
-
Digit/letter disambiguation:
(I've modified the glyphs1and the4: I've removed the serif under 1, and opened the 4; that is not essential for the purpose of disambiguation, it is a matter of personal preference.) -
Latin-Cyrillic-Greek disambiguation (in normal and bold weights):
-
Text samples:
-
Making the difference visible:
Note that a wrong letter in a translation may go undetected by an automatic validator, and make unusable the index entry based on it (e.g. in an e-book).
cd GIT/fixed-9x18/src
make
This shall produce 9x18.pcf.gz (normal weight) and 9x18B.pcf.gz (bold).
The easiest way is to overwrite the 2 homonimous files from the
xfonts-base package in the standard /usr/share/fonts/X11/misc/
directory. Just say
sudo make install
(this is the default).
Or you could prefer to add the new fonts in a system-wide manner, e.g.
put them into /usr/local/share/fonts/fixed; you can do it with
make install whereto="/usr/local/share/fonts/fixed"
Or you could install them locally, just for the current user; the
usual locations are ~/.fonts (obsolete and deprecated) and
~/.local/share/fonts. Hence, say e.g.
make install whereto="~/.local/share/fonts"
The problem of these non-default locations is that not all systems have them in the Xorg font path. You may check it with
xset -q | grep -A1 Font
If the chosen destination is missing in the list, you have to add it
into your ~/.xinitrc or the startup file of your window system:
xset +fp /usr/local/share/fonts/fixed
or
xset +fp ~/.fonts
Hopefully thereafter you should be able to see the installed fonts by the means of
xfd -fn '-Pok-fixed-medium-r-normal--18-120-100-100-c-90-iso10646-1' &
xfontsel -pattern '-Pok-Fixed-*-R-Normal--18-120-100-100-C-90-*-*' &
The fonts are in the Public Domain, as described in the Debian repository (SIL Open Font License, Version 1.1).