Hi AR, On Montag, 4. Februar 2008, a r wrote: > It is indeed an interesting functionality. The following are my > observations that I had after playing with it a bit:
> 1. Other schematic editors do not connect a currently drawn net to > the attracting point by default. Instead, the net follows the cursor > closely (the attracting point is displayed but is not used) and only > if a "snap" key is pressed the net is forced to connect to the > attracting point. It's a bit like having a "monostable" magnetic-mode > option instead of "bistable" one, as it is now. Interesting variant. I thought about adding a modifier key to not connect to the magnetic marker. But the Control key is used for non-orthogonal nets. > 2. It should be possible to merge this patch with the net routing > patch that I've posted to this list a few weeks ago. As far as I can > see the functionality provided by both of them is orthogonal (i.e. > the magnetic mode patch does not attempt to improve routing > algorithm). I bet you will like it. I take a look at it and maybe add a branch to my git repo. > 3. Code searching for a nearest attracting point is another place > where a spatial index could be useful. Yes. If the performance is too bad with the current code. "Premature optimization is the root of all evil". > Finally, there is a number of other possible improvements in the net > drawing code: > a) (matter of preference) I don't like nets being routed from the > current mouse position. This is the default behavior when the net > drawing mode is activated from a keyboard shortcut. I agree that it > may save somebody one click but it is inconsistent with other > activation methods. Same thing applies to other modes as well > (selecting, moving, copying etc.). Is it worth having a configuration > option here? In other tools it is called a "prefix mode" (which is > turned off by default, i.e. used is required to click on the > starting/reference point). I like that because I usually draw from the free space towards the pins. > b) The net is terminated when user clicks on its ending point again > (that's OK). IMHO it should also be terminated on a double-click (at > any location) or if the end point is on an existing net/pin. Mmmh. I like the left-right mouse click combination to finish a net. > c) The net is terminated when its mid-point happens to be placed on > an existing net/pin (and this mid-point is then used as an ending > point). This is very confusing and IMHO should be disabled > (alternatively that connected mid-point could become a new start > point). I've no opinion about that. > d) Net consolidation code. There are some bugs here (try for example > drawing nets on top of an existing net) but IMHO the biggest problem > is that segments should end at nearest net connections, pins or > bends. Currently the whole straight part of the net is being treated > like a single net segment which make editing more difficult and does > not work well in the moving mode. That would require quit a little coding work. Regards Werner _______________________________________________ geda-dev mailing list [email protected] http://www.seul.org/cgi-bin/mailman/listinfo/geda-dev
