Re: blot_diameter

[Han-Wen Nienhuys]

address@hidden writes:
> 1: 256-char limit
> 2: messy selection of fonts
> 

> 3: The font-standart only supports fonts in 16 (IIRC) different heights. This 
> is a problem for the note-heads (having critical heights).

No, 3 is not a problem. Heights are read from the AFM file, not from
the TFM.

> I think that the solution of 1 and 3 is to divide the font - not by
> style but by function. I.e. one font for the note-heads, etc.  2
> could be nicely (IMHO) solved by allowing #'style to be a list
> instead of a single entry. Preferably the first entry is used, but
> for glyphs not present, try the 2nd entry instead, etc.  So to get a
> classical quarter rest, one would do
> 
> \property Staff.Rest \override #'style = #('classical 'default)
> 
> perhaps some sort of macros shuld be used for this:
> 
> classicalRests = #('classical 'default)
> ...
> \property Staff.Rest \override #'style = \classicalRests

Indeed, but this scheme could also be used for the one-font-per-style
approach. I like the one-font-per-style approach better, since it
matches better with how other currently available fonts do it. BTW we
already have a font-series / font-shape field that could be used for
this list-thing.

Or perhaps it is better to have a general fallback mechanism for
fonts, i.e.

 #'font-sequence = #'((font-shape medieval) (font-shape default))

because that would also work for text fonts

 #'font-sequence = #'(((font-shape italic) (font-series caps))
                       (font-series caps))

> In 1.5.something a size-19 font was added. Why?

It has always been there, probably because  musixtex  also had it.

> Is there a good reason why the font contains reduced noteheads
> instead of just using the full-size heads in the smaller font?

no. But I also don't see a good reason to remove them :-)

> > to set ancient notation. This approach would also get rid of a lot of
> > conditional code for setting the glyph style. For now, the ancient
> > font is called parmesan.
> 
> Well, the messy code would be moved from c to mf... :-/

ok, point taken. 

-- 

Han-Wen Nienhuys   |   address@hidden    | http://www.cs.uu.nl/~hanwen/