From: David A. Scocca (scocca@email.unc.edu)
Date: 11 Mar 01, 03:52 EST
From: "David A. Scocca" <scocca@email.unc.edu> Subject: Re: A few issues... --On Saturday, March 10, 2001 9:59 AM -0500 Jon Guyer <jguyer@his.com> wrote: > At 9:34 AM -0500 3/10/01, David A. Scocca wrote: > >> To my mind, the way Alpha works violates three logical principles: > > These are only logical principles if you start with a bunch of > assumptions. Your way works if that's the way you expect it to work, but > it severely constrains your ability to extend a selection. Alpha's way > likewise works if that's the way you expect it to work, but it severely > constrains your ability to contract a selection. As I said in my original post, I think it depends on whether the user views the use of the arrow keys as a single, long action (move the insertion point from location A to location B) or a series of shorter actions (expand the selection one character left, then one row up, then one character left, and so on). >> As I indicate above, mouse selection clearly implements model >> (a)--we care about the endpoints but not the route traversed. > > Does it, now? I suggest you go back and try it again. Alpha's mouse > selection is actually not consistent with either model. I said mouse selection, not using the mouse to _extend_ the selection. Again, the multiple-action versus single-action paradigm: the difference between (a) holding down the shift key, pressing cursor keys, and lifting the shift key is like holding down the mouse button, moving the mouse, and lifting the mouse button and (b) pressing an arrow key with the shift key down is like clicking the mouse with the shift key down. Insight into a possible model for compromise--consider the state when the shift key is pressed. If there is NO selection, then treat the cursor position as an anchor point and have the arrow keys expand or contract the selection from that point. If there IS a selection, have the arrow keys expand the selection in the appropriate direction. This would mean that releasing the shift key and then pressing it again would "lock" the selected block of text so it could no longer be contracted. People who see the sequence of arrow keys as a single "action" will not release the shift while making a selection. People who see each shift-arrow-key combination as a separate action can logically release and re-press the shift (once) and will see they behavior they expect. I think that sort of arrangement would get to the heart of the underlying differences in logical models. Dave * Dave Scocca SAS for the Macintosh Resource Page * * dave_scocca@unc.edu http://metalab.unc.edu/sasmac/ * _______________________________________________ AlphaTcl-developers mailing list AlphaTcl-developers@lists.sourceforge.net http://lists.sourceforge.net/lists/listinfo/alphatcl-developers