incorrect extra ligatures in t1.etx for monowidth fonts?

Lars Hellström Lars.Hellstrom@math.umu.se
Mon, 30 Aug 1999 12:41:09 -0400


Vladimir Volovich wrote:
>t1.etx does not distinguish between monowidth and non-monowidth fonts
>(unlike ot1.etx). As a result, it contains ligatures which are not
>present in monowidth fonts. These are e.g. f-ligatures; their presence
>usually does not cause any trouble because fontinst's etx files
>provide slots with faked F-ligatures.
>
>But there is also one extra ligature, "endash minus -> emdash" which
>is present in roman T1-encoded fonts (ecrm1000) but is absent from
>typewriter fonts (ectt1000). This results in _different typesetting_:
[example snipped]
>Therefore it seems that it is better to make fontinst more accurately
>follow ligature setup of encodings (in particular, even remove
>f-ligatures from typewriter fonts, etc).

I agree something ought to be done about this, but there are a couple of
matters that ought to be sorted out before we can decide exactly what. One
is whether there is a distinction between monowidth and typewriter
fonts---as the ETX files are currently written, ot1.etx produces a
typewriter font iff it is monowidth, whereas t1.etx never produces a
typewriter font.

(It seems to me that things are generally quite a jumble in this area
(encodings in the TeX world), mainly due to various unfortunate early
decisions on how fonts should work with TeX (such as handling the input
encoding problem of converting two hyphens to an endash by a construction
in the output encoding). Perhaps you would like to take a look at Section 1
of CTAN:macros/latex/contrib/ supported/relenc/relenc.tex; it is a more
detailed explanation of my opinions on the subject, although from a LaTeX
view. Comments are welcome.)

Lars Hellström