[OSM-dev] Potlatch node positioning

[Richard Fairhurst]

Robert Funnell wrote:

> W.r.t. the first point in my message of Jun 25: I see that the
> behaviour really did improve in 9c. I've experimented a bit and I find
> that inserting a new point does in fact sometimes cause a shift in the
> existing way. However, the shift is very small, perhaps just one
> pixel, so not a problem. Perhaps it's a rounding issue.

I'd guess so, yes. Previous Potlatch behaviour was that it inserted  
the point at the mouse click. Now it inserts at the nearest point on  
the existing line to the mouse click. So any inaccuracy is probably  
from rounding.

Can't remember if I answered this one:

> Second, sometimes when I try to reposition a node Potlatch decides that I
> didn't move it enough to be worthwhile so it just leaves it where it was.
> This makes it difficult to make fine adjustments. Is this   
> intentional? If so, is it really useful?

It's how Potlatch differentiates between "click" (to select) and  
"drag" (to move). If it's for a short period of time, and not a very  
big move, then Potlatch decides it's a click, so doesn't move the  
node. Getting this one right for everyone has never been at all easy -  
I think I've rewritten the code about 31242 times - and all feedback  
that it's too far to one extreme or the other is helpful. :)

cheers
Richard