[REQ]Remove InputSensor>>mousePoint

Reinier van Loon R.L.J.M.W.van.Loon at inter.nl.net
Tue Apr 6 07:59:07 UTC 1999


Duane, you wrote:


>Hmm... not sure about that.  It is possible that someone might want the
>cursorPoint and the mousePoint to be different things.  If we interpret the
>mousePoint to be the position of the physical mouse,
I can't see why one wants to know the position of the mouse on the mouse
pad.

> and the cursorPoint to be the position of the indicator on the screen,
>then gridding is an example where they might be decoupled.
I would prefer griddedCursorPoint then.

>There are some actual examples of the this is a couple of commercial Mac
>draw/paint applications, and MacApp explicitly supported this idea.  When
>gridding was on, the cursor jumped between grid points to be near the
>mouse.  This gave better feedback to the user about what would happen when
>the button was pressed, since without it it's not as clear which grid point
>would be chosen.
The Morphic developers have chosen to do this at the Morphic side. I agree
that griddedCursorPoint might be a worthwhile addition to InputSensor. OTOH,
one could add a method to Point as in:

    sensor cursorPoint griddedTo: aPoint

where aPoint is a normalized point which defines the grid in both x and y
direction.

All I'm saying is that there are only 11 senders of mousePoint (like
Celeste). I would like those to use cursorPoint and then remove mousePoint.
226 clients of cursorPoint get 3 message sends to obtain the cursor point:

cursorPoint sends mousePoint send primMousePt which is primitive.

I want to keep primMousePt (it's categorized private anyway) and recode
cursorPoint to use the primitive.

That means that 226 clients suddenly have a very short message path for
obtaining the cursor point:

Cheers, Reinier.





More information about the Squeak-dev mailing list