Misplaced code

[Hilmar Schlegel]

Thierry Bouche wrote:
> 
> Concernant « Re: Misplaced code », Alan Jeffrey écrit : «
> » Hmm... I think I agree with your assessment here: the boundarychar is an
> » encoding, not a metrics issue.  This might be something that's so old
> » that it's become a `feature' though...
> 
> i doubt...
>  Isn't it even safer _not_ to set it when it's not used?

There are two aspects here: 
i) to define which character in a font makes up the boundarychar, this
depends of course on the specific encoding of the font
ii) whether or not to include ligatures with the boundary

> This said, as visiblespace and some other ascii charachters are the
> first ones you'd like to replace by usefull glyphs when you face fonts
> with many usefull glyphs (as a Swash italic fonts, with automated
> swash initials/endings through <BDCHR> ligs -- so keeping also the
> normal versions of the whole u&lcs for inner-word letters)  i alway s
> had to move it elsewhere in such bizarre encodings (\char0 was usually
> OK).

I believe it is best when one can define the boundarychar by defining
which character is to be used for it, independent of its special
position in the encoding in question. Therefore no code position but a
character name.

Hilmar Schlegel

###