I'm designing a canvas based software with pickable objects. User interaction is performed by a custom state machine based system. I need to specify the selection process, but I need some recommendations or best practices on how users should pick objects. As a requierement I need to implement selection by LeftMouseDown and multiselection by Ctrl + LeftMouseDown. For example: in Windows 8 OS, selection of desktop icons are performed on LeftMouseDown, but multiselection is performed on LeftMouseUp (I don't know why). Are there any description of the processes or state machines that controls the user interaction with mouse picking of objects on a standard basis?.
There's a pretty thorough chapter on selections in the 1999 "Microsoft® Windows® User Experience", see here.
I don't know if there's a more recent document or if that chapter's material is available somewhere on-line.