Hi,
Currently if you select, say, a rectangle and scale it using a knot, the origin for the scale is taken as the opposite corner of the rectangle's bounding box.
This is fine if you want snap-to-bounding-box behaviour, but it presents problems if you want snap-to-nodes since the node near the opposite corner of the bounding box will drift wrt its original position.
Does anyone have any suggestions as to where to take the origin from? If it needs to be optional, how should we decide which point to use for a given instance?
Thanks
Carl
On Tue, Jun 22, 2004 at 10:51:28AM +0100, Carl Hetherington wrote:
Currently if you select, say, a rectangle and scale it using a knot, the origin for the scale is taken as the opposite corner of the rectangle's bounding box.
This is fine if you want snap-to-bounding-box behaviour, but it presents problems if you want snap-to-nodes since the node near the opposite corner of the bounding box will drift wrt its original position.
(Note I've thought for only a minute about the question.)
I wonder whether it should default to the opposite corner of the bounding box of the _unstroked_ shape, which for an unrotated non-rounded rectangle coincides with the node corner.
The plus sign can be dragged to indicate the scale origin. Perhaps we should do snapping on it, to facilitate having the origin be exactly on a node or (probably harder to program) on a stroke corner (assuming miter joins).
pjrm.
On Tue, 22 Jun 2004, Peter Moulder wrote:
On Tue, Jun 22, 2004 at 10:51:28AM +0100, Carl Hetherington wrote:
Currently if you select, say, a rectangle and scale it using a knot, the origin for the scale is taken as the opposite corner of the rectangle's bounding box.
This is fine if you want snap-to-bounding-box behaviour, but it presents problems if you want snap-to-nodes since the node near the opposite corner of the bounding box will drift wrt its original position.
(Note I've thought for only a minute about the question.)
I wonder whether it should default to the opposite corner of the bounding box of the _unstroked_ shape, which for an unrotated non-rounded rectangle coincides with the node corner.
The plus sign can be dragged to indicate the scale origin. Perhaps we should do snapping on it, to facilitate having the origin be exactly on a node or (probably harder to program) on a stroke corner (assuming miter joins).
Ah yes, I hadn't thought of that. The plus sign seems to have been removed for scale, however --- IIRC it was because it can get in the way of moving small objects. It already snaps to the grid, so that bit is done.
Any other thoughts? Or should it just be a preference?
Carl
Ah yes, I hadn't thought of that. The plus sign seems to have been removed for scale, however --- IIRC it was because it can get in the way of moving small objects.
Yes, and because scaling around a non-standard center is not as frequently needed as rotation around a non-standard center. The center mark is just hidden, but it's there; switch to rotation mode, position it, switch back to scale, and the new position will be used for scaling.
Any other thoughts? Or should it just be a preference?
Which one - to show or not to show center mark? Possible, does anyone else think it's a good idea?
On Tue, 22 Jun 2004, bulia byak wrote:
Ah yes, I hadn't thought of that. The plus sign seems to have been removed for scale, however --- IIRC it was because it can get in the way of moving small objects.
Yes, and because scaling around a non-standard center is not as frequently needed as rotation around a non-standard center. The center mark is just hidden, but it's there; switch to rotation mode, position it, switch back to scale, and the new position will be used for scaling.
Fair enough, I've no problem with that change.
Any other thoughts? Or should it just be a preference?
Which one - to show or not to show center mark? Possible, does anyone else think it's a good idea?
No, I mean where to put the scale origin by default: on the bounding box before or after stroke? In other words, on the bounding box of the shape, or the bounding box of its nodes?
Cheers
Carl
On Tue, 2004-06-22 at 13:05, bulia byak wrote:
Any other thoughts? Or should it just be a preference?
Which one - to show or not to show center mark? Possible, does anyone else think it's a good idea?
We have too many preferences already. :P (IMO, preferences can become a crutch to avoid finding the best default...)
I think the most sensible thing for "scale mode" is to only show the center mark while the shift key is being held down.
That way, users can see it and move it if they are about to do a centered scale, but it will not be in the way otherwise.
-mental
participants (4)
-
bulia byak -
Carl Hetherington -
MenTaLguY -
Peter Moulder