Re: Ties within chords inconsistency

[Simon Albrecht]

Hello David,

pardon if I insist another time.

Am 09.09.2015 um 22:40 schrieb David Kastrup:
> Joram <address@hidden> writes:
>
> > Hi Simon,
> >
> > I see. The difference is: I would not have expected that to work. But I
> > can see the benefit if it would, now.
> Sigh.  It works perfectly.

Of course it works as long as you keep inside the current way of 
distinguishing in-chord and whole-chord ties. But, as I already said 
earlier this day, ‘this means reconsidering the relation and the 
handling of in-chord and whole-chord ties’. I do not doubt your 
authority in interpreting the way LilyPond thinks, but I have been using 
LilyPond for four years now and invested no small amount of time into 
understanding how it works. My point is that we shouldn’t be insisting 
on striking internal logics too much, if any user who has not delved 
into the exact internal workflow of the program will encounter problems 
with code that is perfectly sensible /on the surface/.
To return to our particular topic: the distinction between ‘ties inside 
simultaneous music expressions within one voice which will be arranged 
into chords only later’ and ‘ties within chords’ is very subtle and far 
away to the musician(*). I can unfortunately not speak about technical 
feasibility, or rather difficulty, but how about the following approach:
Any tie which is not attached to a chord in the <>~ manner is 
interpreted as in-chord tie. This would shift the preference from 
whole-chord tie to in-chord tie for ‘doubtful’ cases like a~ (as opposed 
to obvious cases like <a~> or <a>~), so to speak. I can’t think of any 
cases where that would again cause problems(**), and it seems like a 
more robust approach. After all, user-friendliness is not to be despised :-)

What do you think?

Yours, Simon

(*) And although professional programmers probably form a major part of 
the LilyPond community (?) I think they shouldn’t be our main target group.
(**) Nobody would expect the tie in
\new Voice << { c''~ c'' } { a' a' } >>
to mean a whole-chord tie, that is two ties for both c and a.