Conference bulova::decw_jan-89_to_nov-90

You accidentally pressed CTRL-SHIFT-F3 that enables "mouseless mode".  Press
it again to turn it off.  Note that the Wait light is turned on while
in this mode.

Or do a DIR/TITLE=ARROW and see lots of other notes on this topic.

		Steve

RE: The various folks that tripped over <CTRL/F3> (including myself):

Is there a way (in a future release of DECwindows) to get some component
of DECwindows to generate a box that says "you just pressed <CTRL/F3> and
are now using the arrow keys for mouse operations..." and then provides an
explaination of the way back out of this mode? 

How about just popping up a graphic of a gun, going off, pointed at a foot?
(Or, better yet, at a mouse.)

There *is* an indication - the "Wait" light on the LK201 is lit when in 
pseudo-mouse mode. But that's obviously a compromise (instead of "wait", 
it should be labeled with a picture of a mouse with its little feet in the
air).

I'm just full of good ideas today...

    
    Can't use Mickey(tm), Disney will get mad.  Try Mortimer Mouse :-)
    
    

    re: .3
    
    Problem is, that when the next generation keyboard starts shipping,
    it is minus the wait light and the compose lamp.
    
    I agree, the system should put up some sort of ICON (ear-con, nose-con)
    to indicate that the user, using all of his thumbs, unknowingly
    put the machine into pseudo mouse mode. Just think if this was done,
    we wouldn't see "How come my arrow keys move the mouse" entries in
    this conference that appear every so often.
    
    Jeff                                  

        How about changing the pointer to        A
        one with the 'universal symbol'          |
        for movement by cursor keys ...      <--   -->
                                                 |
        (I do recommend better graphics          V
	though ... )

	Kevin

RE: .6 (the nifty arrow icon in dead-rodent mode)

	Cute idea, but would it be obvious enough?  (and how
	would I know from the arrow icon what to do to get out
	of this mode?)

	Is it possible to do an informational pop-up ("Yo, you 
	told me the rodent was dead!  Press <CTRL/F3>..." :-) if
	in dead-rodent mode and any rodent motion is detected?


RE: whoever commented on cutting down on the "Hey, the arrow keys..."
comments around here...

	You got it.  And for the customers -- the folks who don't
	get to see any of the same "Hey, the arrow keys..." notes.

>>>	Cute idea, but would it be obvious enough?  (and how
>>>	would I know from the arrow icon what to do to get out
>>>	of this mode?)

    Okay. How about a popup box that says "Do you really want a dead mouse?",
    with "Okay" and "Cancel" button (which you must select with the pointer
    ... )

    yug ug ug ug ug!

    Kevin

    Have a cat come out and eat the pointer, replaced by a figure of a key
    with arrows pointing nsew.
    
    
    
    		- Andy Leslie

I figure if the mouse is really dead, the user probably knows about the arrow 
keys and perhaps how they got to be mouse substitutes.  If not, the user 
probably doesn't need arrow keys at all, since they are largely dead in the 
water.

The real problem is when the mouse is alive, and the arrow keys are duplicating
the mouse, instead of being regular arrow keys.  In this case, the user might
think of looking at some customization menus to figure out if they set up some
option incorrectly.  Might it be a possibility to have the Keyboard 
customization window contain an option to set the arrow keys =arrows/=mouse?

This would not replace the CTRL/F3, but augment it.  


							JP

    Would it be possible to have a *real* mouse operation (i.e. clicking
    a button) cancel the 'dead mouse' mode?  Even without any knowledge of
    this feature, it's quite likely a user would poke the mouse before
    assuming the system was 'broken'...

    It might be useful to look at some history here.  The primary reason
    that the mouse emulation is in the VMS X-windows driver is because it
    was in our UIS driver.  The reason it was in our UIS driver was because
    if the mouse died, you couldn't even log in, and CSSE demanded we fix
    the problem in the name of servicability.
    
    It would appear to anyone reading this conference that this "feature"
    causes more problems than it resolves in DECwindows.
    
    It sounds like it may be time to retire this feature.  For the 7 people
    that really do use it out there, add some bizarre SYSGEN parameter
    to turn it back on (bit 7 in TTY_DIALTYPE is free, right?).
                                                               

    Hmmm.  It seems to me that if anything DECwindows is even more
    useless without a mouse.  You might be able to log in, but that's
    about all you can do with a DECwindows system and no mouse.  Maybe
    Motif will fix this.  It also turns out to be a fairly reliable way
    of doing "fine" positioning.
    
    I think that the complaint is in the lame mechanism chosen to get
    into/out of the mode and the lack of feedback that tells the user
    that he is in that mode.  Both CTRL-F2 and CTRL-F3 were complained
    about before DECwindows ever shipped as likely to be among the
    top-10 "user errors" (along with compose-space).  There is a distinct
    lack of complaint about this feature on VWS - because it is modeless -
    the user must *actively* make this happen via CTRL-SHIFT-EDIT_KEY
    

We didn't follow the modeless model of VWS because we knew that things like
extend selection (which require a shift key) couldn't be done if the shift
and ctrl key had to be held down to get into the mode.  There might have
also been some problems with the driver and the server both trying to deal
with keyboards that made this very hard (impossible?) to do.

And yes, the theory behind Motif is that it can be mouseless. A fairly silly
requirement that comes from early days of MS-Windows running on IBM-PCs
where they couldn't assume a mouse.  However, there are some people (and
markets) that can't get along with mice, so I suppose it's a good thing.

Another bit of history is that the old square mice that we used on the VS/I
had a dead on arrival rate of about 1 in 2 early on.  I've never seen one
of these round mice go bad.  I'm sure that it happens, but I suspect the
number of people that have been bagged by getting into pseudo-mouse mode and
are confused by that is MUCH greater than the number of people that
have been helped by it.

    
    Can't "nuke" features and have happy customers.  While it might cut
    down on the QAR's, it's a drastic way of fixing a bad human interface
    decision.  The best you could have done is not offered it in the first
    place.  It's too late now to 'nuke it' - you've got to replace it or
    fix it.
    
    You can, as Jeff suggests, make it harder to get to (CTRL-SHIFT-F3),
    devise a more positive visual indication that the keys are in mouse
    mode (a new cursor, a pop-up on the screen, etc.), or make it modeless
    so that it only operates when a user is actively making it happen -
    like say CTRL-F3-EDIT_KEY.  You can put a timeout on it - say if there
    is no movement in a certain period of time the mode reverts - or if
    a non-edit_key is pressed the mode reverts.
    

My current theory is that it should be harder to get into AND it should 
auto-cancel with real mouse input.  Seems like that would solve most/all of
the outstanding problems.

The only question I have is whether anyone who currently USES this arrow-mice
would yell and scream about auto-cancel.

Burns

    Don't forget about very sick (but not dead) mice that are giving
    junk.  This is a case where random mouse input shouldn't turn
    off the arrow-mouse.
    
    Also becareful about making it too hard to get into.  All of these
    CNTRL-SHIFT-F3 while turning your head to the left and coughing
    reminds me of my "user-friendly" Macintosh where hitting 5 shift
    keys in a row will do something...  After a while it becomes
    black magic passed down from generation to generation.  IMHO,
    DECwindows already has too many "hidden" key click/press sequences.
    
    				-John

    How about these new server logicals that cause our
    servers to do the appropriate SETMODE QIO:
    
    DECW$AUTOCAN_MOUSE_EMULATION
    
    	TRUE	- Mouse movement cancels mouse emulation (default)
    
    	FALSE	- Mouse movement and mouse emulation can
    		  interact without canceling.
    
    DECW$DISABLE_MOUSE_EMULATION
    
    	TRUE	- CTRL-SHIFT-F3 is not intercepted by the drivers.
    
    	FALSE	- CTRL-SHIFT-F3 enters mouse emulation mode (default)
    
    This preserves the ability to use mouse emulation, makes it harder to
    get into and makes it easier to get out of.  It also allows the user
    to override things.
    

    re: .17
    Entering pseudo-mouse mode does not disconnect the real mouse.  So a
    a sick (babbling) mouse would probably interfere with your keyboard 
    'mouse' anyway.

    I agree completely with .17...
    
    Get into pseudo-mouse mode and, simultaneously, use cursor keys and
    move the mouse about.  You will find that BOTH affect the direction of
    the mouse.
    
    The best (prob'ly ONLY) way to get ONLY pseudo-mouse mode is to put the
    terminal into that mode and unplug the mouse.
    
    And, FYI, even NEW DEC products require the pseudo-mouse emulation!
    
    \burt
    
    (one_who_implemented_pseudo_mouse_mode_on_an_unannounced_X_terminal)
    
    P.S.   It ain't a mouse anywayz.... it's a PUCK!  :-) :-) :-)

I love these long threads about small arcane features.  Allow me to add
my two clicks!

If your mouse babbles, unplug it!  Then swing it by it's tail around
and around---and then FLING IT!  You'd really be surprised how many
cubes away it will go.  You will feel much better! 

Make real mouse input or motion cancel the F3 and the problem is solved.

    
    Being the one to fix this, I chose the easiest way out.
    From now (V3) on, clicking on a REAL mouse button disables
    the pseudo mouse feature and the WAIT light goes off.
    So, clicking on the real mouse is equivalent to keying in CTRL-F3.
    Jeff, I admit that this doesn't solve the WAIT-less problem on
    LK401's but that's another can of worms.
    
    
    
    Eva.  

    re .22:  Sounds like a satisfactory compromise.  Well, I've gone and
    customized my keyboard to make my operator window key Shift-F1 just to
    get away from Ctrl-F3 in two directions.  (Now instead I'll get
    confused when I use other people's workstations.)

    --Simon

>    Being the one to fix this, I chose the easiest way out.
>    From now (V3) on, clicking on a REAL mouse button disables
>    the pseudo mouse feature and the WAIT light goes off.

(Boy you just can't satisfy some people! ;-) ((namely me)))

Why didn't you make mouse motion also cancel this mode?  From my experience
writing server code I'd bet this should be just as easy.....

I'll tell you why I think it should:  For those users (like me) who use
window manager technology different from the Dec sypplied DWM, you don't
click as frequently since focus follows the pointer...  Also I believe
that a confused user is more likely to move the mouse than click when
he/she accidentally gets into this mode.

So..... Can I petition to have the motion cancel this also??

Regards....

    re .24:  Actually, the keyboard-mouse mode, if it's used at all, is
    significantly enhanced by the mouse for traversing the screen.  If you
    make mouse movement cancel the keyboard mouse, it'd make the keyboard-
    mouse mode almost useless.  Supplemented by the mouse, it can sometimes
    be useful.

    --Simon

> Actually, the keyboard-mouse mode, if it's used at all, is
> significantly enhanced by the mouse for traversing the screen.  If you
> make mouse movement cancel the keyboard mouse, it'd make the keyboard-
> mouse mode almost useless. 

I agree that the method you mention is usefull: Move mouse, fine tune
with arrow keys.  This would be quite usefull in a CAD application, if
that application lacks `object snapping' capabilities. 

I disagree with your assessment that cancelling the mode with mouse
movement makes it useless.  After all it is only a simple CTRL-F3
away, and that is not much different from using a shift key,
especially when you form the habit.  The technique would then become:
Move mouse, Ctrl-f3, fine tune with arrows.  And I think the motion
cancel would greatly relieve the confused user.

However, these things are greatly personal.  You can't please everybody;
witness the plethora of keyboards.... 

I believe that the effected solution does not relieve the purported
problem, namely user confusion, at least not as well as the motion
cancel method.