
On 2007-August-08 , at 23:42 , bulia byak wrote:
On 8/8/07, jiho <jo.irisson@...400...> wrote:
Some minor quirks:
- the cursor (orange circle) seems to leave some traces behind it
sometimes. maybe this is related to my version of GTK (2.10.12)
No, this is just our renderer. Hopefully a switch to cairo will fix this.
- the cursor does not disappear when it is moved outside the drawing
area (typically to adjust some settings)
Yes, I'll look into this
great!
another small possible quirk: I find it difficult to have the tool push things in diagonals. I noticed this when using a (regular, i.e. not optical) mouse with a medium-complex path (B&W traced bitmap). The difficulty in moving in diagonals comes from the mouse, in my opinion, but is made visible by the tool. Indeed it starts pushing things apparently immediately, in the direction of the first mouse movement (which is often horizontal due to the position of the hand on the mouse when I click and maybe also to the fact that the movements of the ball inside are better detected horizontally and vertically than diagonally). Then I instinctively correct the movement but, due to the computing time I imagine, there is a lag in before the change of direction is visible and, when it is visible (which is not the case in most situations), it is often abrupt because I corrected too much. With a stylus, things are usually quicker and more precise but the fact that inkscape starts pushing things quickly can also be a problem sometimes. Indeed I use a stylus like a regular pen (i.e. it draws only when I touch the tablet -- as opposed to: it is tracked when touching the tablet and draws when the button is pushed). In this case, when touching the tablet, the stylus usually moved a little bit in the direction opposed to the direction is it tilted toward (i.e. it moves left a bit because I am right handed and hence hold my stylus tilted to the right). This can result in some unwanted pushes.
Could there be some black magic introduced, such a delay before starting to push and a computation of the mean displacement vector during this delay. And then maybe a computation of the mean displacement vector using a moving average window to avoid sudden changes in direction. Maybe those things (initial delay and size of moving average window) could be made sensible to machine computing power for them to feel right (but I imagine this must be complicated...). Anyway, I thought I would report my impressions... and the tool is really great no mater what! Thanks again.
JiHO --- http://jo.irisson.free.fr/