Re: Setting X-offset and Y-offset graphically.

[Janek Warchoł]

2013/4/23 Richard Shann <address@hidden>:
> On Tue, 2013-04-23 at 17:16 +0200, Urs Liska wrote:
>> David Nalesnik <address@hidden> schrieb:
>> >I'm a little confused here too.  I watched the demo (very cool, BTW!)
>> >and you are starting with a default curve, if I'm not mistaken.
>
> Well, only if the user clicks accurately on the control points marked
> with the red crosses. The user has set up the coordinate system with the
> first cross-hair click on the center staff line, and then the rest is
> relative to that.
>
>> >Couldn'tthe
>> >results of dragging the various control points be expressed as
>> >displacements to the curve you started with, rather than as absolute
>> >values?  Then adjustments would be made automatically with a
>> >change in layout.
>>
>> That's what I thought initially too.
>> But maybe this default curve isn't Lilypond's? Then he'd be out of luck.
>
> What would be needed is to get the user to click on the control points
> accurately and then work out the differences from there. That could be
> done...

What about this: ask the user to drag the control points.  Then he
doesn't have to click precisely on the control point.
In more detail:
- user enters slur editing mode, control points appear.
- user clicks near first control point
- holding the mouse button down, he moves the cursor 2 units
horizontally and 1 unit vertically, then he releases the button.
- this becomes \shape #'((2 . 1) ... )
- similarly for other control points.

This way it doesn't matter how precisely he clicked on the control
point, because we don't have to set up any absolute coordinate system
- we just measure the displacement he did when mouse button was held
down.  All is relative!
Actually, the same mechanism could be applied to moving grobs you
showed before.  Things should become way simpler and more precise, and
won't even need showing refpoints.

Does this make things clearer?
Janek