Re: Concepts that may be missing (or are at least hidden) in LilyPond

[Urs Liska]

Am 27.04.2015 um 15:00 schrieb Carl Sorensen:
> On 4/26/15 10:55 PM, "Andrew Bernard" <address@hidden> wrote:
>
> > In what way are measures missing? Half the world calls them Œbars¹ by the
> > way - that¹s what they are under in the index. Are you simply referring
> > to terminology in the documentation?
> LilyPond creates bars (the things that divide measures) just fine.
> LilyPond has a Timing engraver that keeps track of the current time within
> a measure (so you get warnings if you place a bar check that doesn't
> coincide with the end of a measure).
>
> But if you look, for example, at the Music Encoding Initiative (MEI), they
> have a container called the measure, which contains musical elements over
> a certain period of time[1]. So notes are placed into measures.
> Presumably this then allows a text parser to search through the data
> schema, find measures, and eliminate them.
>
> LilyPond has no data structure for a measure.  The only timing
> relationships in the LilyPond model are sequential (one note follows a
> previous note) and simultaneous (two notes start at the same time).  The
> full timing of the score is worked out by Iterators when the score is
> produced.
>
> There are some advantages to the LilyPond way of doing it.  LilyPond
> easily supports different time signatures in different staffs.  Music is
> easily copied, because it has no absolute timing information.
>
> There are also disadvantages to the LilyPond way.  It is common for people
> to think things like "the crescendo should cover measures 2-4".  There is
> no way to express this kind of relationship in LilyPond.
>
> I'm not convinced that LilyPond should add a "measure" object to its
> language.

I would strongly oppose to *replacing* the current approach with a 
container based approach (I don't think you are even considering that, I 
just want to make that clear). Apart from the polymetrics stuff the 
container  imposes significant limitations with regard to items crossing 
the measure boundaries. Or (in the case of Finale and Sibelius at least) 
with incomplete measures.
AFAICS from blog post etc. it is really painful to talk one of these 
programs into writing tuplets over bars (maybe even beams over 
barlines?), and when you try to edit existing music and delete something 
from a measure you can't really leave that measure without the program 
complaining (or doing stupid things).

For a concept of "measures 2-4" we already have a powerful technology 
right before us: the edition-engraver. It is not fully mature and ready 
to be included in LilyPond itself. But apart from providing means to 
completely separate content from separation it is also good at injecting 
stuff "measure-wise".

For example right now it is possible to write:
\editionModList full-score Score
  \break
  #'(7 16 22 (27 2/4) 30)
to apply line breaks (or whatever) at this list of measures (including 
partial measures)

What would be good to have is a notion of "select everything in measures 
2-4 - over all possible parts" as this is something that is currently 
more or less impossible.

Urs

> But I am convinced that LilyPond really doesn't have the
> concept of a "measure" object in its schema right now.
>
> Thanks,
>
> Carl
>
> 1. http://music-encoding.org/documentation/guidelines2013/cmn#cmnMeasures
>
>
>
> _______________________________________________
> lilypond-user mailing list
> address@hidden
> https://lists.gnu.org/mailman/listinfo/lilypond-user