Some comments on the ask for fixed-width measures / absolute proportional notation.
Hopefully, this will help clarify what the actual ask is.
One solution would seem to involve increasing the modularity of
LilyPond, to enable folks to author their own versions of the underlying
systems. In this way, an alternative layout/spacing engine would be
able to provide truly strict proportional alignment of grobs given
whatever constraints are useful to modern composers.
Lilypond already has several different spacing options.
Perhaps more options would be better, but I think that this basic concept is already baked into the cake.
So, this does not seem inherently problematic from a conceptual point of view.
But such added
modularity greatly increases the complexities of managing compatibility
with other systems. Imagine all of the snippets we have so far and how
they might be making assumptions on internal mechanisms.
What are the "other systems" with which we'd need to manage compatibility with?
I'd imagine that most snippets use and work with the default spacing, and would continue to do so if additional options were added.
And snippets that work with non-default spacing are already specifying that spacing, and so would also continue to work.
Why would adding a new spacing option be a problem for snipptes?
All I have ever asked is that the measure layout engine be able have its
nice variable barline position attribute turned off when people want
fixed length bars.
I am well aware of the massive code and architectural
issues this may present and it appears to be impossible. I find this
curious since the hallmark of lilypond is flexibility and extensibiilty,
yet it has a doggedly unchangeable viewpoint of only one way to layout
bars in the page.
I suspect the reason has more to do with the point that David K. alluded to, which is that the concept of actually proportional notation is somewhat non-sensical:
the width of each symbol is non-zero, yet there is no direct relationship between a symbol's width, and the amount of musical time it takes up. In fact, some symbols take up space but no time, like accidentals.
So, it will not in general be possible to fit an arbitrary array of musical symbols into a given space and have them both be readable (non-overlapping) and have their horizontal layout reflect duration.
While your ask is to set the barline spacing to a fixed width and not a direct call for absolute-proportional notation, I suspect that in effect these amount to the same thing, insofar as they pose similar problems:
* If you want exact widths, you will have to allow for overlapping symbols.
* If you don't want overlapping symbols, then you can't get exact measure widths.
It may be that your use cases don't incur these issues.
But, a general layout engine will need to deal with such cases.
Yes, I have read the essay on why liypond is superior
because it varies bar position in repetitive music from line to line,
with the justification that this enables musicians to find their place
on the page more easily should they look away and then back again. It's
a nice idea, but speaking a as a player myself, I have never accepted
this or bought that idea. I don't have that problem, neither does any
player I have met. Where is the actual evidence for that? I think that's
really a purely aesthetic decision. And the thing is, it is a decision
forced on us by lilypond and the point is it is not what everybody wants.
I agree that justification "[variations in] bar position in repetitive music from line to line [that] enables musicians to find their place on the page more easily" is bogus. So, let's toss that out. But that is hardly the motivation for flexible layout; rather, it is just the result of the general set of rules.
However, I think it is safe to say that literate musicians in general will not be confused by this practice in the slightest, so the results don't really need a justification. I get that some people think they want some kind of fixed spacing. But that is not a good reason to say that the default behavior is problematic. Will any musician have difficulty reading the standard flexible spacing? What are the problems?
When I say that some people "think they want it", what I mean is that, until they accurately address how the rest of the layout should work, this ask is not really fully thought out, so it is not a clear ask. They want the benefits, but have not considered the problems, or at least, have not adequately conveyed how those problems should be addressed. Maybe in the use cases that have come your way, these issues don't crop up. But, a spacing engine will have to have an answer for more complicated cases in general, and "give me fixed with measures" doesn't provide enough guidance on how to deal with those cases. (And because the solution will most likely be to invoke some form of flexible spacing.)
Forcing a fixed with for some measures means that everything else is not a fixed wiidth, so it becomes a derived width based on the line length and the remaining space, regardless of what that content is. So, the complete algorithm will have to lay out those bars using a different algorithm, since in general the remaining material cannot be guaranteed to take up the same amount of musical time as would be inferred from the available horizontal space.
If the main point of proportional notation is to coordinate temporal flow with horizontal alignment, then this may be in contradiction with asking for fixed width bars, in the general case.
It seems to me that a lot of jazz players want chord sheets with lined
up barlines for _ease_ of reading, and that request comes up from time
to time on this list. Another user case.
Andrew
I certainly understand that many people want something like "always have 4 bars per line". That is common enough, and it is possible to do that already.
Believe me, I understand the mindset, as someone who for years did Jazz parts by hand with pencil and paper, one of the first things you do is grab a ruler and mark all the barlines.
When the content is consistent from bar to bar, then the spacing is actually going to be even, so this will be achieved by default.
And when the content from bar to bar differs considerably, I suspect that in most of these cases, people would actually prefer flexible spacing, rather than overlapping symbols. Maybe their first thought is that evenly spaced bars is best, but once you have too many notes to that space, I have difficultly believing that anyone would prefer overlapping symbols to stretching one bar to accommodate the glyphs, and shortening the others.
So, I don't think we should put this common ask of N-bars-per-line in the category of fixed-width bars, since when push comes to shove, I don't think that is actually what most people want, even if that is their presumed starting point.
Thanks,
Elaine Alt
415 . 341 .4954 "Confusion is highly underrated"
address@hidden
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-