bulia byak wrote: 
On Thu, Jul 31, 2008 at 12:49 PM, Diederik van Lierop
<mail@...2312...> wrote:
  
If we know where the clipboard data was originally located (I don't know
if that information is stored)
    

Of course it is stored - the objecty is copied with its own transform
and other attributes, and that determines its position. When we "paste
in place", we simply paste and nothing else; on regular paste, we
paste and then move the pasted stuff to the point under cursor.

OK, thanks for the info.

This also answers your other question on snapping: paste should snap
exactly the same as if the original object, without any copy/paste,
was just dragged by mouse to the current location in Selector.
When pasting, it used to snap only to grids and not to other objects or to guides (as it would when dragging). So it's a little bit different.

But if we only snap to grids, then it would make more sense to me if we'd use a multiple of the grid-pitch, instead of really snapping to the grid. When making technical drawings I always end up with some object aligned at half the grid pitch. Usually I don't want to double the grid density for that single object. When pasting a selection with both aligned and unaligned objects, I'd like to keep the aligned objects aligned and not snap that single unaligned object to the grid. Besides, a rectangle might have all its corners aligned but it's center can be in between grid lines, so when this snaps the corners will become unaligned.

So, what's against using the multiple of the grid pitch? It does have some advantages IMHO.

Diederik