Re: Lily and petrucci (whaddapair!)

[William R. Brohinksy]

Break a leg!

And when you get back, a few responses...

Mats Bengtsson wrote:
> I'm just to take off to Norway to play string quartet for a week
> and don't have time for a long answer. However, from a quick
> look at your wish list, it seems that most of this is already
> available or can easily be implemented using a few lines of
> inline Guile code. For the bar lines, take a look at the
> properties barNonAuto, timing and barAlways (combined with
> defaultBarType = ""). Note that Lilypond only will consider
> line breaks where there's a bar line (that's why barAlways=##t
> may be useful).
> 

I hope this isn't a really dumb response (Note: This is a Non-Truth. I'm
really good at dumb responses, especially like this one!) but I don't
know the first thing about Guile, and haven't a clue about where to find
out more. Would one of you folk who have been down the path be willing
to show me where the beginning of it is? (further Note: I kind of told
another fib, becuase I was able to copy lines of Guile or Scheme or
whatever out of ancient-font.ly and make a new file in which I tried
some stuff, and some of it worked, and some of it failed spectacularly.
However, I really know so little about how to deal with Guile (or maybe
Scheme?) that I couldn't really begin to do anything about it.

Another aspect of this: I've found the Internals part of the manual very
difficult to use because it's all online (on the internet) and I have a
very slow connection. I love the fact that I can print out the users
manual and take it with me to study when I'm places that don't have
computer connections at all (like Boy Scout Camp a couple of days next
week). Is it possible to come up with the internals manual in ps form?


> Changing the durations of notes can be done using Guile code,
> take a look at input/test/smart-transpose.ly and
> input/test/reverse-music.ly to get some inspiration.
> 

Yup. Looks like LISP to me. (Y'know: Lots of Infuriating, Spurious
Parentheses?) Why can't any of these things ever have a nice FORTH
internal language? (oh. sorry. Flogging my prejudices again!) Well,
anyway, without getting a grasp of Guile, and how it's interfaced with
(apparently Scheme? and) Lilypond, I am but a babe in the woods. It
needn't stay this way, but after my attempt to find something that just
illustrates how Scheme code is supposed to look like, I'm discouraged
enough. (I'd probably have a different attitude if the last 2 months
hadn't found me learning 2 new IDE's and 3 languages, of which I've used
two for a very small part of programs I had to make, and will probably
never see again.)

> For the clef position, you can get whatever you want, see
> the reference manual.

I've tried some of this stuff. Either it's my total ignorance about
Scheme and Guile, or something doesn't work quite right, because none of
it worked for me. On the other hand, ancient-fonts.ly did, and I wonder,
do we really need four different ways to specify a clef, or would it be
good to have one fairly complete and uniform system that will allow
specifying a clef, cleftype, clef position and octivation, regardless of
modern, vatican, mensural, or whatever, and then leave really fancy
things (reduced-sized clefs, additive clefs (like the hufnagel clef that
shows both the stylized f and c on the same staff) for coding? My
request is for the ancient clefs to work like the modern ones do (as
well as some mechanism for adding other clef forms, although I'd be
perfectly happy if just adding a petrucci set was all that got
accomplished for now.)

> 
> For the number of staff lines, have you looked at
> input/regression/number-staff-lines.ly?
> 

Not before now. Would you check the output of ancient-fonts.ly and see
if you get four lines for the top staff? number-staff-lines.ly does make
a single 3-line staff just fine. But ancient-fonts turned out two fives,
as I said, on my machine.

> To change the layout of note heads and other symbols, some
> Metafont hacking is needed. For the square symbols, this
> should be easy even if you don't know Metafont. Take a look
> at mf/feta-bolletjes.mf and play around with the parameters.
> Erase all old feta20.tfm and feta20.*pk files and use
> tex testfont
> to get a table over the feta20 font.

::grin:: OK, I'll admit it. I haven't done any metafont hacking for
years. (15 of them to be honest) And I wasn't that good at it then. I've
never been that cool in visual arts. But that is a very effective way of
throwing the gauntlet back in my square, so I guess I'll dig out my
MetaFontBook and start thwacking.

Query: is it best to hack Petrucci fonts into the feta fonts, and if so,
which characters would be best to use? Or should I hack a new font? How
would I get my characters properly linked into lilypond? hack it myself
without external direction and then try to get Jens or whoever to accept
the changes? I'd feel better at this point to have some direction.
> 
> I hope someone else can fill in more details and the questions
> I didn't cover.
> 
>   /Mats

I'll admit that my wishlist was proffered from the hope that the people
who already understand the internal workings of Lilypond and Scheme and
Guile and such might be interested in implementing them, rather than as
a list of things I was prepared to do in my copious (uh huh) spare time.
But if that's what it will take, I certainly do have 'something to do'
with it, so this would be a good time to start learning what I need to
about scheme and guile and the gorey internals of lilypond.

Does anyone know how to squeeze 38 hours out of each day?

raybro