Re: [Groff] more on ellipsis...

[Werner LEMBERG]

> > [Ideally, typing `?...' and `!...' should be sufficient, for example
> >  using ligatures:
> > 
> >    ?... -> \[?..]
> >    !... -> \[!..]
> >    ...  -> \[...]
> >    etc.
> > 
> >  Unfortunately, groff doesn't support this (yet).]
> 
> That is a very good idea which I've also thought of proposing to
> the mailing list.  It shouldn't be too hard, either.  (I hope.
> I havn't looked at the code yet.)

It is a major nontrivial task.  I invite you to work on it.

> But groff already has a mechanism for recognizing the fi, fl,
> etc. sequences in the input stream, and a builtin mapping to the
> corresponding ligature characters, which "only" has to be extended
> (maybe dynamically from the manuscript as well, on a per-font
> basis).

Correct.  The static tables should be made dynamic.

> Of course, the font description files will have to be slightly
> modified (compatibility issues?), for example
> 
>   name AGaramond-Regular
>   internalname AGaramond-Regular
>   spacewidth 250
>   encoding text.enc
>   ligatures
>   fi fi
>   fl fl
>   ff ff
>   ffi Fi
>   ffl Fl
>   kernpairs
>   A z 6
>   A w -42
>   ...
>   charset
>   ha      500,626 2       0000    -- asciicircum
>   ti      500,297 0       0001    -- asciitilde
>   ...

Yes, this is one part of the solution.  In addition, we need a `liga'
request to add and remove ligatures -- some languages (like Turkish or
Portuguese) don't use ligatures resp. only particular ones.

I don't see compatibility issues if we introduce a new keyword, say,
`ligtable'.


    Werner