Graham Percival <gra...@percival-music.ca> writes: > On Wed, Nov 10, 2010 at 01:18:49PM +0100, Werner LEMBERG wrote: >> > I would like to state that I consider that a step in the wrong >> > direction. If we make Lilypond smart enough to figure out what is >> > wrong, we should use that smartness for doing the right thing >> > instead of educating the user. >> >> Hmm. You mean that a user should never need to specify the context of >> a grob, and lilypond should be able to automatically walk over all >> contexts to find the right one? This sounds useful.
More like "if a user specifies a context (whether explicitly, or implicitly by using the current one) not supporting a particular grob/property, then Lilypond will walk through the parents until finding one that does it". Or put differently: "where a context does not support a particular grob/property on its own, it inherits the respective grobs from its parenting context". > It's not possible in all cases -- consider trying to set the staff > name to "Violins" for the staffgroup containing violin 1 and > violin 2 -- but it could be done for most cases. The real > question is whether it's worth doing for "most" cases, and whether > the automatic method might increase confusion. The automatic method has the advantage that we can introduce layers of rather lightweight contexts without affecting existing semantics. Chords can be formed in a subcontext of their own, and manipulated accordingly. Subvoices can be used for tasks now done using multiple versions pasted over each other using transparency and disabling collision detection. I mean, ugh. -- David Kastrup _______________________________________________ bug-lilypond mailing list bug-lilypond@gnu.org http://lists.gnu.org/mailman/listinfo/bug-lilypond
