[texhax] A sign-change puzzle

[Uwe Lück]

Pierre,

At 04:28 12.12.06, pierre.mackay wrote:

>I hope you have seen my response to Phil's message to me, which I sent out 
>to TeXhax.

Indeed, I have ...

>I have rewritten the macro, and tested it on a reasonably complex 
>example.  Even so, it
>will necessarily require a sound understanding of \expandafter and 
>\noexpand to
>avoid hidden traps.  But for any of the uses I am likely to make of it, it 
>works fine.
>If you didn't see the new version, let me know, and I will send it to you.

... although I haven't studied the details ...

>Your offering is vastly more sophisticated than anything I would try.  I 
>am afraid I am
>too lazy to attempt to collect the font pair-kerns and use them in the way 
>you suggest.

The package I referred to doesn't require the user to "collect"
the kerns. Rather, it *measures* the implicit kerns automatically.
The user only decides on how much they should contribute
to the kerning finally.

>May I confess that, once one is committed to "stealing sheep" as the old 
>Scotsman Goudy
>termed it, I am not sure that the eye will be able to pick up the slight 
>variants in
>uppercase pair kerning that you are taking care of.

Well, I experienced strong prefences with examples ("VORWORT")
-- however, my most recent impression is that it depends ...

>    And surely none of us would
>risk being haunted by Goudy for letterspacing lowercase characters 
>(except, of course, in Greek).

I have used my mechanism for "part" front pages only indeed,
so I am not much afraid ... (and I know almost nothing about Greek).

All the best,

Uwe.