Folks,
please calm down. Nobody here is insinuating anything! As far as I
can see, we have a clash of concepts that is probably not resolvable
in *any* satisfying way. Consequently, the only way forward is to
minimize the fallout, namely by providing explanations what a musical
term 'foo' means in LilyPond, and why there are differences to its
standard, colloquial usage.
> Most helpful of all was the suggestion that the grob descriptions
> given as IR 3.1 ...
>
> https://lilypond.org/doc/v2.24/Documentation/internals/all-layout-objects
>
> ... be expanded to list *all* properties for a given grob, rather
> than only the properties that a grob *changes* from interface
> defaults.
Jean has shown a possible solution to that, at least for the HTML
page; I'm attaching his image here (taken from
https://gitlab.com/lilypond/lilypond/-/issues/6210). We now have to
find a way to implement that in Texinfo while still getting decent
results in the other output formats.
Now answering to other comments.
Matthew wrote:
> The point isn't to list properties with no effect, but to list *all*
> of the properties that *do* have an effect
As demonstrated by Harm, this is not possible, for various, mostly
technical, reasons. In addition to that, the IR is a static document
that *cannot* reflect the state of LilyPond at an arbitrary point of
time while processing an input document. Instead, it presents
LilyPond in a pristine state, before any `*.ly` files has been loaded,
more or less (including LilyPond's own startup files).
> But one might well also ask, if there are useless properties with no
> effect, then why are there useless properties with no effect,
Again, Harm demonstrated that this can be dynamically changed in many
cases, suddenly 'activating' properties and vice versa. If there are
grobs that LilyPond really ignores by using hard-coded values instead,
it should be documented, and it is a bug if there is no documentation
for that.
Hi Werner,
Thanks so much for summarizing the thread; super helpful. My point in revivifying the discussion from 2021 was to upvote the prototype that Jean provided back then, so thanks for re-linking to his example.
I am a touch worried about something, though:
Matthew
> The point isn't to list properties with no effect, but to list *all*
> of the properties that *do* have an effect
Werner:
> As demonstrated by Harm, this is not possible, for various, mostly
> technical, reasons.
Is what's being communicated here that (for technical reasons) users won't *ever* be able to visit a single page of documentation somewhere and see all the properties that affect a grob?