Re: \set vs \override

[Han-Wen Nienhuys]

On Tue, Nov 24, 2009 at 3:50 AM, David Kastrup <address@hidden> wrote:

> Sigh.  I guess I give up.  Yes, I understood that.  Pretty much from the
> get-go, and also from the manual.  The unanswered question is _why_ you
> want only _one_ of the two different things happen to _one_ half of the
> properties, and the _other_ of the two different things happen to the
> _other_.

Because that's what people want?

  \override Script #'direction = #UP

is useful.

  \set Script = #(blah)

is ridiculous, because it would overwrite

   (Script
     . (
        (cross-staff . ,ly:script-interface::calc-cross-staff)
        (direction . ,ly:script-interface::calc-direction)
        (font-encoding . fetaMusic)
        (positioning-done . ,ly:script-interface::calc-positioning-done)
        (side-axis . ,Y)

        ;; padding set in script definitions.
        (staff-padding . 0.25)

        (stencil . ,ly:script-interface::print)
        (X-offset . ,script-interface::calc-x-offset)
        (Y-offset . ,ly:side-position-interface::y-aligned-side)
        (meta . ((class . Item)
                 (interfaces . (font-interface
                                script-interface
                                side-position-interface))))))

thereby removing all of its functionality, including appearance of the
symbol in the output.

If this is so unlogical to you that it needs to be explained, I give
up explaining things to you.

>> At some point we had \set Foo.Bar \override #'x = #y syntax for this,
>> but it was deemed to confusing, so we gave it a different syntax.
>
> I'm fine with the two different syntaxes for the two different actions.
> But why is one action only for context properties, the other only for
> grob properties (which are also pre-registered in the context)?
>
> --
> David Kastrup
>



-- 
Han-Wen Nienhuys - address@hidden - http://www.xs4all.nl/~hanwen