Re: [9fans] mouse vs key

["Douglas A. Gwyn" <DAGwyn@null.net>]

Boyd Roberts wrote:
> i do remember thinking (when i was using the blit) that
> i wish that damn code would give me back my mouse!  then
> you'd make a massive overcorrection and the mouse cursor
> would be anywhere except where you wanted it to be.

The way we implemented cursor hysteresis (using a tablet pen
rather than a mouse) in the PEAC system, for example in
locking onto menu items, was simply to track the actual
locator device coordinates all the time but "snap" the
displayed cursor location to whatever grid once we were
within lock range of a grid position (e.g. something like
1/2 of an item width for menus).  That is much better than
adjusting the locator coordinates to always match the jumped
cursor location.  Our users were very happy with the
behavior: solid visual feedback when an item became
clickable, a "dead space" around the exact grid location
within which small locator motions would not change what
was clicked on, and continuous locator mapping into the
display space so that one's motor intentions were always
consistent with observed behavior.