Frank

>Lars
>
> > Also, perhaps I should point out that what Frank called glyph collection
> > seems to be pretty much what I call encoding in relenc.tex, and what he
> > called encoding seems to be pretty much what I call coding scheme in
> > relenc.tex.
>
>yes indeed. do you agree that our names are better?
>
No I do not; my intention with that paragraph was simply to give other
people some first aid in sorting out things.

>point is what we call "glyph collection" and you "encoding" is a set ie
>something unordered (a collection) which is why i think calling it encoding is
>confusing the issues as in my book encoding means associate a mapping with a
>set. right?

Firstly, the reason I started calling it encoding is that when I started to
write relenc (that was almost two years ago), I thought about it as a
package for use under LaTeX2e. In LaTeX2e, that NFSS axis is called
encoding; it's that simple. Secondly, I do think about encoding as being a
mapping---a mapping from some set containing letter tokens, "other" tokens,
some LaTeX commands, and certain patterns of such to "typeset output" (a
vague term, I know, but I think you get the idea). I suppose your glyph
collections are the ranges of my encodings, but I do not think that glyph
collections are the objects that one should choose with NFSS3. It does
matter how things get mapped---you cannot simply be content with that there
is something which gets mapped to what you want, you must also know what
this something is.

>example: the glyphs in a ps font (ie those in the AFM file some of which are
>compositive) form a collection (you call it encoding) and the encoding vector
    As used here, I do not. -------->~~~~~~~~~~~~~~~~ /LH
>defines an encoding (ie the subset + its mapping to numbers) that can be
>actually used (you call it coding scheme)

Thirdly, the encoding mapping has a natural decomposition in two parts,
namely what is done in TeX and what is done after TeX. I use the term
coding scheme to describe the combination of all mapping that is done after
TeX. Your description seems a bit odd---the coding scheme maps numbers to
glyphs, not the other way round.

Lars Hellström