On 11/07/14 08:01, Urs Liska wrote:
Am 11.07.2014 07:00, schrieb David Kastrup:
tisimst <address@hidden> writes:
All,
Is there anyone who is VERY against distributing music fonts in
binary form
(i.e., as otf, svg, etc.files)? I just don't see how we can make
other music
fonts available by forcing them to have a metafont source file. I
guess that
could be nice, but it seems like so much work to do that. I have
about 4 or
5 alternate music fonts that people could use and I certainly don't
want to
convert them to metafont. They are currently designed and built with
fontforge.
What do you think?
Spirit of the GPL is delivering source code, defined as "preferred form
of modification, including all scripts etc". Now fonts are reasonably
separate anyway, but that's what we should stick with. METAFONT is just
one possibility here.
If the fonts are derived from some upstream source, automating the
derivation as much as possible makes sense in order to facilitate
integrating future improvements from upstream.
IIRC Abraham uses scripts to bring existing fonts (e.g. Bravura) to
the usable form. If that should be a completely automated process and
would for example make it possible to update the Profondo font
automatically if a new version of Bravura comes out that would be
quite good. Bravura itself is *not* delivered as source code, for
example.
Urs
I guess I am missing something here.
Why do we need other fonts as part of LilyPond anyway? In that I can
already use (can I not?) any font that I have installed with the
appropriate markup command, I can see why emmentaler and feta were
included as that was what Han and Jan 'created' when they created
LilyPond and I can also understand why we have users that extend the
glyphs (all the weird and wonderful arrows and dots and shapes for
noteheads etc.), after all LilyPond needs at least one font to base its
code on.
So what is the point of including more *as part of the code base*?
This isn't meant as any kind of argument against or questioning of the
request to add another font, I just don't understand what the purpose
would be, other than 'because we can'.
My only concern is that we then need to include these new fonts (if
applicable) as part of the regression test suite surely? And make sure
that LilyPond's spacing calculations are not going to be compromised by
the addition of new fonts and that it isn't going to start to create
more exceptions because one font's dynamic 'f' happens to be wider or
fatter than LilyPond's (if you see what I am getting at).
Regards
James