Re: [PATCH] Re: syntax change for \cresc

[Reinhold Kainhofer]

Am Samstag, 24. April 2010 11:50:20 schrieben Sie:
> Hi Reinhold, Graham,
> 
> 2010/4/23 Graham Percival <address@hidden>:
> > As far as I'm concerned, yes.  Maybe wait 24 hours in case anybody
> > objects?  But then go ahead.
> > 
> > For anybody skimming this: the patch includes a convert-ly rule
> > which changes any current \cresc into \deprecatedcresc, so this
> > change doesn't require any manual attention to ly files.
> 
> Sorry I missed the 24 hours objection deadline...

No problem, I've been busy preparing my OrchestralLily presentation, anyway.

> I'm not a programmer but if I understand well Reinhold's proposal this
> new \cresc command would be to text cresc *with spanners*, right?
> (the equivalent to the current \crescTextCresc c2\< )

Exactly. The old cresc did also this exact thing, with the only exception that 
it applied to the next note instead of the previous one.


> I would have preferred to have cresc *without spanners* as default
> behaviour for the \cresc (and so \dim) command.
>
> There are a few reason of this, the main one being that it's like that
> I'm already redefining the \cresc command for now  ;D
>   cresc = #(make-dynamic-script
>     (markup #:line (#:left-align
>                     #:normal-text
>                     #:italic "cresc.")))

You can continue using that definition, of course.

However, your approach has the problem(s) that (1) it only creates something 
that looks like a crescendo, but lilypond does not know this internally, so 
(2) MIDI won't be affected at all, and (3) should we ever get a MusicXML 
backend (or any converter to any other format), it will also not properly 
export the crescendo.

The \cresc, \dim, \decresc commands need to produce spanners internally, 
otherwise lilypond won't have that information available. Lilypond does not 
necessarily have to print a spanner line, though.

> But I see other reasons why \cresc should be used for the one without
> extender line:
> 
>   – cresc without spanners is far most common in the scores I have
>    (and in the ones I write);

You know that you can (globally or locally) override the spanner type and even 
hide the line?
    \override DynamicTextSpanner #'dash-period = #-1.0

It's just a question whether this should be made the default or not...

>   – the behaviour of the previous \cresc command was without spanners;

No, \cresc created a dashed spanner.
\version "2.12.0"
\relative c' {
  \cresc d1\mf d1 d1  d1\!\f
}

>   – from what I remember from the previous discussion, it was said that
>     \cresc , \dim , \decresc , ... would produce text (de)cresc without
>     spanners by default.  And for that with spanners a syntax like
>     \<"cresc. poco a poco" would be used.

That won't work without modifying the parser and complicating the syntax. I 
don't think we have han-wen's support for this.

AFAIR, there was no consensus about the exact output, so what you remember was 
a statement by one person, not a general agreement.

>     But I think something like \crescTextSpan "cresc. poco a poco"
>     should be fine too as a syntax (and for custom cresc text witout
>     span we could use \crescText "cresc. blabla" for example).

See input/regression/dynamics-custom-text-spanner-postfix.ly for how to create 
dynamic text spanners with arbitrary texts.

> In every case the command should be inserted *after* the note, as
> everybody agreed I think (and as Reihold's proposal).
> I hate these commands that requires to place something before the note!

Me too.


Cheers,
Reinhold
 
-- 
------------------------------------------------------------------
Reinhold Kainhofer, address@hidden, http://reinhold.kainhofer.com/
 * Financial & Actuarial Math., Vienna Univ. of Technology, Austria
 * http://www.fam.tuwien.ac.at/, DVR: 0005886
 * LilyPond, Music typesetting, http://www.lilypond.org