Re: gEDA-user: Strange user interface behavior with gschem-1.6.2.20110115

[Josef Wolf <jw@xxxxxxxxxxxxx>]

On Mon, Sep 05, 2011 at 09:13:06PM +0200, Kai-Martin Knaak wrote:
> Josef Wolf wrote:
> 
> > So this feature can save at most one click? 
> 
> It provides a means to reliably hit that tiny spot of a pin point,
> while zoomed out.

But it prevents me from drawing the connection and connects to totally
different net.

> The feature is "magnetic-mode". it is not "auto-complete".

I don't care how it is called. As long as it connects to anything other than 
the object that had the circle at the time of the click, it is mis-behaving.

> > At the expense of using the undo functionality fairly often?
> 
> Well, I can't recall to have entered your szenario, ever. So I certainly
> do not use undo faily often, because of it.

Honestly, I have tried to make the desired connection by locating the cursor
to different positions in respect to the positions of the target and of the
nets around it. Even tried to do variations of x/y distance and such. I was
not able to make the proper connection.

In this simple case, the conflict could have even be resolved by walking the
y-axis before the x-axis, but I was not able to trigger this. It looks as
if this tool always goes the x-axis first.

> > In fact, I was trying to find a
> > way to draw the connection in question for several hours before I
> > posted to the list.
> 
> For the archive: The "correct" way is to provide one or more additional
> node with a click somewhere on the empty canvas.

Yes, tried that also. But I had to go farther away to prevent the circle to
appear at the wrong pin. And when I was far enough, it again connected to the
wrong net.

> If gschem insists on
> a poin, where you don't want the net to go (hinted by the circle), then 
> hold the [ctrl] key and the circle will always stay with the mouse cursor.

OK.

> > IMHO, this works against user expectation. After all, the whole point
> > of the circle is to signal the user where the connection would be
> > done. Connecting to anything other is not exactly what the user would
> > expect. 
> 
> Fair enough.
>
> > From the perspective of the user it is totally irrelevant
> > what algorithm is used internally.
> 
> But it is relevant to know, what this mode is intended for -- magnetic-
> mode, not auto-complete. I am sure, you got the point now :-)

I don't care how it is called as long as it connects to the wrong
destination. Oh, I already wrote this above ;-)

> >> (Auto routing for nets would be a great, nerdy feature :-)
> > 
> > I tend to disagree here. It works against user expectation. It
> > _would_ be a useful feature _if_ it would be disabled automatically
> > in situations where it would connect to something different than the
> > point where the circle is shown.
> 
> IMHO, silent denial of service would make matters worse.

How come silent denial of service be worse than silently connecting to the
wrong net?

Please note that I never wrote that it should be a silent denial of
service. The circle should simply not appear as long as the path to the circle
is blocked by some object. I don't see a DoS here..

> It would 
> create a sense of uneasiness and loss of control -- "sometimes, the
> tool works, sometimes it goes on strike...". A pop-up warning would
> be the better choice. Text might be something along the lines 
>   "Could not connect with a straight L"

Uh, no. The circle should only appear when the path to it is not blocked.
No reason for anybody to expect some tool to work as long as no circle is
shown.


_______________________________________________
geda-user mailing list
geda-user@xxxxxxxxxxxxxx
http://www.seul.org/cgi-bin/mailman/listinfo/geda-user