steffen@ihuxa.UUCP (Joe Steffen) (01/29/85)
When you press a mouse button to pop-up a menu, should the top item be the default selection, or the last item you selected on the menu? The latter has the advantage that clicking the menu button repeats the last operation, but 1) is error prone because you have to remember the last operation and if you forget you will select the wrong one, 2) you have to decide whether to move the mouse up or down to get to the correct item if you don't want the last item you selected. Thus you will save time if you repeat a selection and lose the half second deciding which way to go if you don't repeat. The Teletype 5620 DMD(TM) Terminal under the layers program defaults to repeating the selection by clicking, but this can be changed by a program running under layers. Should a program stick to the default so as not to confuse the user? Would novices prefer one way and experts another? What do other mouse-based systems do? -- Joe Steffen, AT&T Bell Labs, Naperville, IL, (312) 979-5381
ksbszabo@wateng.UUCP (Kevin Szabo) (01/30/85)
In article <722@ihuxa.UUCP> steffen@ihuxa.UUCP (Joe Steffen) writes: >When you press a mouse button to pop-up a menu, should the top item be the >default selection, or the last item you selected on the menu? The latter >has the advantage that clicking the menu button repeats the last operation, >but > 1) is error prone because you have to remember the last operation > and if you forget you will select the wrong one, > > 2) you have to decide whether to move the mouse up or down to get > to the correct item if you don't want the last item you selected. I don't know what other systems do, but if you have a Menu based system the underlying motivation is that you NEVER have to remember. Thus if you pick (1) and use the last operation you should highlight the operation as if you had manually selected the light-button. Same goes if the default is the first item in the menu; that item should be highlighted so the user knows what is coming. My personal feeling is that the user should be able to define the menu contents and the default case. They should also be able to chose whether the default case is the last operation chosen. If they haven't altered the menu the default should be the operation that is used the most. The user should have the choice to tailor their environment. Of course, if there is a standard way for your system to do things by all means stay consistent. Kevin -- Kevin Szabo watmath!wateng!ksbszabo (U of Waterloo VLSI Group, Waterloo Ont.)
judd@tove.UUCP (Judd Rogers) (01/31/85)
The default action for a menu poped up via a mouse button press should be to
do NOTHING. What if your finger slips off the button and the last thing you
selected is some sort of delete?!
The Macintosh and Xerox Star system do it this way.
For novices this is best as it is least confusing. but for experts...?
A command from the keyboard might be better for an expert (Macintosh supports
this).
--
Spoken: Judd Rogers
Arpa: judd.umcp-cs@CSNet-relay
Uucp:...{allegra,seismo}!umcp-cs!juddnorman@sdcsla.UUCP (Donald A. Norman) (01/31/85)
In article <722@ihuxa.UUCP> steffen@ihuxa.UUCP (Joe Steffen) writes: >When you press a mouse button to pop-up a menu, should the top item be the >default selection, or the last item you selected on the menu? The latter >has the advantage that clicking the menu button repeats the last operation, An important principle, but one little remarked upon, is that of spatial consistency. Spatial information is a very important and powerful cue for remembering things. It is important not to violate this principle. A pop-up menu that reogranizes itself with each usage violates this principle. (I am almost tempted to say that it violates human decency, but that, of course, would be flaming.) There is a second alternative, followed by some Xerox PARC pop-up menus: the ordering of the menus stays the same, but the initial mouse position (i.e., the highlighted item on the menu) initially appears on the last item selected with that menu. Even here, I am not sure whether this violates the principle, because even though the ordering of items on the menu remains constant, the relative movement of the mouse to pick an item changes from usage to usage. An experiment is in order. An alternative that appears to combine the best of all features (except perhaps adding some confusion on first experience) is for the inital mouse pisition always to be on the menu header (a non-item), but that the highlighted item initially be the last item selected. Then, a click gives the previously selected item. To make a new selection, move the mouse to the position: any movement into the menu region cancels the initial default. In this case, both ordering of the menu items and relative motions of the mouse remain constant for each use. Donald A. Norman Institute for Cognitive Science C-015 University of California, San Diego La Jolla, California 92093 UNIX: {ucbvax,decvax}!sdcsvax!sdcsla!norman OR ARPA: norman@nprdc
warner@orca.UUCP (Ken Warner) (02/03/85)
Keywords:pop-up,menu The debate on pop-up menus (pums) is a symptom of a fundamental flaw in current computer systems. Not enough screen space to hold a complete work space. No more need be said about the analogy of the tty screen to a desk top. Or the attempts to mimic a desk top with a clever interface. The use of pums to increase the informational capacity of the electronic work space (a 24 by 80 character screen) only over loads a users mental work space. Not only does a user have to concentrate on the task at hand. But also keep reinforcing short term memory information about the contents and use of his current pums configuration. Granted, a statically configured menu system (one that never changes) could be used with less over loading once learned to the point of automaticity. Of course this requires no change in the task at hand. My point is pums wouldn't be needed if there was enough electronic workspace to have all necessary menus and the task at hand visible at all times. So where does this extra workspace come from? Well, there are now many companies producing several types of flat panel displays with the pixel capacity equal to or greater than the currently standard 24 by 80 character screen. For the most part they are low power consumption devices. They are also not very expensive and will be getting even cheaper. I see systems of the future (and not too distant) having multiple screens. Perhaps as many as you want. With very high resolution and color. A user will spread them out on desk or floor, prop them up on books, hold them in his hand or on his lap. Small ones for scratch pads. Large ones for composition. Current window packages already provide the software methods to manage multiple screens. It will be only a matter of time before some daring young company breaks out of the single screen mentality. It will be like breaking a dam. Ken Warner... somewhere in the northwest.
steffen@ihuxa.UUCP (Joe Steffen) (02/05/85)
>When you press a mouse button to pop-up a menu, should the top item be the >default selection, or the last item you selected on the menu? The latter >has the advantage that clicking the menu button repeats the last operation, I can see there is some confusion due to the wording of my original message. I did not mean that the menu would be reordered. I meant that if you press a mouse button to pop-up a menu and selected the third item, the next time you popped-up the same menu the mouse pointer would be on the third item. Thus the last selection is remembered and can be reselected without looking at the menu by just clicking the mouse button. -- Joe Steffen, AT&T Bell Labs, Naperville, IL, (312) 979-5381
marie@harvard.ARPA (Marie desJardins) (02/05/85)
Ken Warner: > My point is pums wouldn't be needed if there was enough electronic > workspace to have all necessary menus and the task at hand visible at > all times. I'd prefer to have a workspace that's reasonably compact, sitting right in front of me. If I want to change the task at hand, I have no objection to pulling down the menu -- it's much more convenient than hunting around a desk for a file folder, for example. I have some objections to being forced to use the pull-down menus for simple commands (this happens a lot on the macintosh) -- I'd much prefer to be able to use keyboard commands when I want to. But I do think that the compactness is one of the beauties of these systems. (Not that a somewhat bigger screen wouldn't be nicer, but you must admit it's easier to keep yourself organized when you're somewhat limited.) Marie desJardins marie@harvard
bobr@zeus.UUCP (Robert Reed) (02/05/85)
Ken Warner claims that the only purpose of pop up menus (pums) is to conserve screen space. He overlooks the fact that pums are also used to reduce arm motion in menu selection. I've seen several workstations that have greater than 24 x 80 character resolution that still use pums for that purpose. Greater resolution (or multiple screens) do NOT eliminate the usefulness of pums. Admittedly, there are other mechanisms for reducing arm motion in menu systems. But given that you are doing menu selection via pointing device, you can either make the menu transient (popup) or permanent (dedicate a porstion of the screen). In either case, it will require a button on the pointing device. You can either overload a button by forcing arm motion to a particular field on the screen (static menus and Apple pull-down menus), or you can dedicate a button for menu use (smallTalk-80, Sun, Apollo, etc.) If you pick a fixed area to hold a dedicated menu, you are limited to the number of entries you can place there, which should be limited to avoid confusion. You could change its contents (when?), or you could have several fixed areas. This second choice opens the question of how the user would know which one was active. By using a popup, you reinforce the model of particular functions associated with particular areas of the screen (windows, objects and other fields), and you trade short term memory use and hand training for recognition and reaction time. I've seen some systems that take advantage of both models, dedicating one button for popups and another for fixed menus. Contrary to Ken's belief about the future, multiple screens have already been tried and generally are a second-best alternative to a single large screen with more resolution. With multiple screens (which are always going to be a more expensive solution than a single screen, whatever the cost of the display), the designer is faced with the problem of letting the user know where to look, and potentially the more complicated interface addressing the multiple screens. The best way to deal with multiple screens would be to treat them both physically and logically as a single, larger screen. -- Robert Reed, Logic Design Systems Division, tektronix!teklds!bobr
davidst@tekgvs.UUCP (David Stubbs) (02/05/85)
Enough of this introspection!
Since this is net.cog-ENGINEERING, I'm enclosing a bit of REAL DATA on
pop-up menus collected from my smalltalk system. Maybe you would like
to develop some reductions or better formats for data reporting or
controls for an experiment that I could run for you. Maybe we could
get some results to publish here instead of speculation. Eh?
David D. Stubbs
Tektronix
Electronic Systems Laboratory
uucp: {ucbvax,decvax,ihnp4,uw-beaver}!tektronix!tekgvs!davidst
USPS: Tektronix, Inc.
Electronic Systems Laboratory, MS 50-383
PO Box 500
Beaverton OR 97077
Phone: 503-627-2627
--------
My current pop-up menu system provides either the previously selected
item or the item selected when the cursor was moved out of the menu
(ie, when the selection process was canceled) as the "Item highlighted
upon menu pop-up".
The first two data items represent the item initially selected and
highlighted and the system time when the pop-up was called for. The
second two items represent the selection and the system clock at
selection time. The time required to data log the first pair is
included in the overall time (maybe we could change this). When "Item
Selected" = 0, I moved out of the pop-up, canceling the transaction.
The menu selections represent a variety of tasks including text editing
and window manupulation.
Item highlighted Millisecond Item Millisecond
upon menu pop-up Clock Selected Clock upon selection
7 1986967 2 1988967
2 1990717 2 1991333
2 1992550 2 1993150
6 1994800 6 1995417
1 2012767 0 2015100
1 2015667 1 2016450
1 2022567 3 2023967
3 2031317 4 2032267
1 2051183 1 2052117
1 2085050 1 2086583
1 2101833 1 2102400
1 2109133 0 2110550
1 2111067 3 2113600
3 2118167 2 2119167
2 2125217 2 2125750
2 2143333 7 2144600
7 2153333 7 2153933
7 2157117 7 2157683
8 2191850 4 2193600
4 2197200 4 2198017
4 2198700 8 2199817
--
David D. Stubbsmauney@ncsu.UUCP (Jon Mauney) (02/06/85)
Why not organize the system so that every menu has an extra entry. The 0th entry in the menu is always the default, but its content changes; it will duplicate one of the other entries in the menu, containing the last selection (or a default chosen by some other algorithm). In this way it is easy to repeat the last selection, but the positions of the entries don't change, and neither does the initial position of the cursor. -- *** REPLACE THIS LINE WITH YOUR SIGNATURE *** Jon Mauney mcnc!ncsu!mauney C.S. Dept, North Carolina State University
rcb@rti-sel.UUCP (Random) (02/06/85)
In article <778@sdcsla.UUCP> norman@sdcsla.UUCP (Donald A. Norman) writes: >In article <722@ihuxa.UUCP> steffen@ihuxa.UUCP (Joe Steffen) writes: >>When you press a mouse button to pop-up a menu, should the top item be the >>default selection, or the last item you selected on the menu? The latter >>has the advantage that clicking the menu button repeats the last operation, > >An important principle, but one little remarked upon, is that of spatial >consistency. Spatial information is a very important and powerful cue for >remembering things. It is important not to violate this principle. A >pop-up menu that reogranizes itself with each usage violates this principle. >(I am almost tempted to say that it violates human decency, but that, of >course, would be flaming.) How about this. the menu has all the selections normally plus a blank entry at the top (where the mouse initially points). Then after each usage, the last used entry is duplicated in the default slot. Random Research Triangle Institute ...!mcnc!rti-sel!rcb
henry@utzoo.UUCP (Henry Spencer) (02/07/85)
> My point is pums wouldn't be needed if there was enough electronic > workspace to have all necessary menus and the task at hand visible at > all times. Of course they wouldn't be needed, but this doesn't mean they wouldn't be desirable. The ability to have something pop up only when needed, and go away (without having to be explicitly put away) when not needed, is a new degree of flexibility that just doesn't exist with papers on a desk top. Personally, I think that this technique is valuable even with a large screen, for clutter reduction if nothing else. The objective at hand is not simulation of a desk top, but provision of a good working environment. The "papers on a desk" model is a useful guide, but it's not the be-all and end-all of environments. One of the big differences between the screen and the desk top is that the screen is under computer control and can be changed -- in useful ways, one hopes -- at high speed. Discarding pop-up menus and related methods just because the screen gets bigger is throwing the baby out with the bath water. -- Henry Spencer @ U of Toronto Zoology {allegra,ihnp4,linus,decvax}!utzoo!henry
gig@ritcv.UUCP (gordon ) (02/07/85)
> Keywords:pop-up,menu > > > The debate on pop-up menus (pums) is a symptom of a fundamental flaw in > current computer systems. Not enough screen space to hold a complete > work space. No more need be said about the analogy of the tty screen to > a desk top. Or the attempts to mimic a desk top with a clever interface... > My point is pums wouldn't be needed if there was enough electronic > workspace to have all necessary menus and the task at hand visible at > all times.... > Current window packages already provide the software methods to manage > multiple screens. It will be only a matter of time before some daring > young company breaks out of the single screen mentality. It will be > like breaking a dam. > > Ken Warner... somewhere in the northwest. Computer systems designed for clearly defined tasks such as CAD/CAM often use multiple display devices already but more general purpose systems face a paradigmatic problem. Multiple displays seems analogous to replacing a database with file cards pasted all over the walls like wall paper. The window or pum is an abstraction that actually simplifies and improves the quality of interaction by reducing complexity of interaction. Or did I miss your point. It is not conceivable to me that I would want to clutter my working environment with displays. That would not be using a system to manage complexity. Gordon Goodman...Rochester Institute of Technology
warner@orca.UUCP (Ken Warner) (02/08/85)
In article <365@harvard.ARPA> marie@harvard.ARPA (Marie desJardins) writes: >.... -- it's much more convenient than hunting around a desk >for a file folder, In one sense this is the point I am trying to make. Using pums is sort of like hunting around a small desk. On a big desk you can spread things out and find what you're looking for faster. >.... But I do think that the compactness is one of the beauties >of these systems. (Not that a somewhat bigger screen wouldn't >be nicer, but you must admit it's easier to keep yourself >organized when you're somewhat limited.) > marie@harvard Keeping organized wears me out. Ken Warner
drforsey@watcgl.UUCP (Dave Forsey) (02/09/85)
In article <455@zeus.UUCP> bobr@zeus.UUCP (Robert Reed) writes: >Ken Warner claims that the only purpose of pop up menus (pums) is to >conserve screen space. He overlooks the fact that pums are also used to >reduce arm motion in menu selection. > This effect: of working in one area of the screen and having to constantly change your focus of attention to another part of the screen, or another region entirely, has been given a name. It is referred to as The Wimbleton Effect. Courtesy Marceli Wien, National Research Council. Ottawa, Canada.
mauney@ncsu.UUCP (Jon Mauney) (02/09/85)
> >.... -- it's much more convenient than hunting around a desk > >for a file folder, > In one sense this is the point I am trying to make. Using pums is sort > of like hunting around a small desk. On a big desk you can spread things > out and find what you're looking for faster. > > Keeping organized wears me out. > Of course, you can see what will happen in 5 years: Everyone will be dependent on overlaid windows on high-resolution screens for all their activities. The result will be *lots* of windows open at once, so that it is very hard to find the one you're looking for. Someone will invent a system with a "window finder": all you have to do is type in the name of the window, and that window moves to the top. If there is no such window, one is created for the corresponding function. Soon, someone will realize that windows aren't necessary, because one can just as easily type the name every time. That's called progress. -- *** REPLACE THIS LINE WITH YOUR SIGNATURE *** Jon Mauney mcnc!ncsu!mauney C.S. Dept, North Carolina State University
riks@athena.UUCP (Rik Smoody) (02/14/85)
> > >.... -- it's much more convenient than hunting around a desk > > >for a file folder, > > In one sense this is the point I am trying to make. Using pums is sort > > of like hunting around a small desk. On a big desk you can spread things > > out and find what you're looking for faster. > > > > Keeping organized wears me out. > > > Of course, you can see what will happen in 5 years: > Everyone will be dependent on overlaid windows on high-resolution > screens for all their activities. The result will be *lots* of > windows open at once, so that it is very hard to find the one you're > looking for. Someone will invent a system with a "window finder": > all you have to do is type in the name of the window, and that > window moves to the top. If there is no such window, one is created > for the corresponding function. Soon, someone will realize > that windows aren't necessary, because one can just as easily > type the name every time. > That's called progress. > -- > > Jon Mauney mcnc!ncsu!mauney C.S. Dept, North Carolina State University Doesn't happen that way in practice. The folks at the labs have their individual styles of laying out their windows juast as they have their styles of spreading out their desks. There are some fastidious folks who have just one paper on their desks at any given time, the rest in file drawers. Their screens look ridiculously neat. Just one or two windows open at atime, the rest hiding in the corner. Then there are us normal folks: a bunch of windows overlapping every which way. So, what's the problem remembering that your "rogue" window is over on the right and the "vi" window is over on the left? It's just like journals on the desk. Rik Smoody, Tektronix