Mode (user interface)
In user interface design, a mode is a distinct setting within a computer program or any physical machine interface, in which the same user input will produce perceived results different from those that it would in other settings. The best-known modal interface components are probably the Caps lock and Insert keys on the standard computer keyboard, both of which put the user's typing into a different mode after being pressed, then return it to the regular mode after being re-pressed.
"An human-machine interface is modal with respect to a given gesture when (1) the current state of the interface is not the user's locus of attention and (2) the interface will execute one among several different responses to the gesture, depending on the system's current state." (Page 42).
In Raskin's sense and according to his definition, an interface is not modal as long as the user is fully aware of its current state. Raskin refers to this as "locus of attention" (from the Latin word locus meaning "place" or "location"). Typically a user is aware of a system state if the state change was purposefully initiated by the user, or if the system gives some strong signals to notify the user of the state change in the place where interaction occurs. If the user's locus of attention changes to a different area, the state of the interface may then represent a mode since the user is no longer aware of it.
Several examples of well-known software have been described as modal and/or using interface modes:
- Text editors – typically are in insert mode by default but can be toggled in and out of overtype mode by pressing the Insert key.
- vi – has one mode for inserting text, and a separate mode for entering commands. There is also an "ex" mode for issuing more complex commands (e.g. search and replace). Under normal circumstances, the editor automatically returns to the previous mode after a command has been issued; however, it is possible to permanently move into this mode using Shift-Q.
- Emacs – has the concept of "prefix keys", which trigger a modal state by pressing the control key plus a letter key. Emacs then waits for additional keypresses that complete a keybinding. This differs from vi in that the mode always ends as soon as the command is called (when the sequence of key presses that activates it is completed). Emacs also has many "major and minor" modes that change the available commands, and may be automatically invoked based on file type to more easily edit files of that type. Emacs modes are not restricted to editing text files; modes exist for file browsing, web browsing, IRC and email and their interaction patterns are equivalent to application software within the Emacs environment. Modes are written in Emacs Lisp, and all modes may not be included with all versions.
- Cisco IOS – certain commands are executed in a "command mode".
- Tools chosen from a palette in photo-editing and drawing applications are examples of a modal interface. Some advanced image editors have a feature where the same tools can be accessed nonmodally by a keypress, and remain active as long as the key is held down. Releasing the key returns the interface to the modal tool activated by the palette.
- Video games can use game modes as a mechanic to enhance gameplay.
- Modal windows block all workflow in the top-level program until the modal window is closed.
Modes are often frowned upon in interface design because they are likely to produce mode errors when the user forgets what state the interface is in, performs an action that is appropriate to a different mode, and gets an unexpected and undesired response. A mode error can be quite startling and disorienting as the user copes with the sudden violation of his or her user expectations.
Problems occur if a change in the system state happens unnoticed (initiated by the system, or by another person, such as the user who was previously using the machine), or if after some time the user forgets about the state change. Another typical problem is a sudden change of state that interrupts a user's activity, such as focus stealing. In such a situation it can easily happen that the user does some operations with the old state in mind, while the brain has not yet fully processed the signals indicating the state change.
A very frustrating type of modality is created by a mode where the user does not find a way out, in other words, where they cannot find how to restore the previous system state.
Examples of mode errors
- The most common source of mode errors may be the Caps Lock key. Other common modes available in PC keyboards are Insert and the other lock keys, Num lock and Scroll lock. Dead keys for diacritics also create a short-term mode, at least if they don't provide visual feedback that the next typed character will be modified.
- PC users whose language is not based on the Latin alphabet commonly have to interact using two different keyboard layouts: a local one and QWERTY. This gives rise to mode errors linked to the current keyboard layout: quite often, the synchronization of "current layout" mode between the human and the interface is lost, and text is typed in a layout which is not the intended one, producing meaningless text and confusion. Keyboard keys in UI elements like "(y/n)" can have opposite effect if a program is translated.
- A frequent example is the sudden appearance of a modal error dialog in an application while the user is typing, known as focus stealing; the user expects the typed text to be introduced into a text field, but the unexpected dialog may discard all the input, or may interpret some keystrokes (like "Y" for "yes" and "N" for "no") in a way that the user did not intend.
- The Unix text editor vi can be notoriously difficult for beginners precisely because it uses modes, and because earlier versions configured mode indication to be turned off by default.
- In many computer video games, the keyboard is used both for controlling the game and typing messages. A user may forget they are in "typing mode" as they attempt to react to something sudden in the game and find the controls unresponsive (and instead their text bar full of the command keys pressed).
In transportation accidents
- According to the NTSB, one of the factors contributing to Asiana Airlines Flight 214 crash was "the complexities of the autothrottle and autopilot flight director systems … which increased the likelihood of mode error".
- Mode confusion was part of the events that led to the loss of Air France Flight 447 in 2009, and the loss of life of 228 people. The pilots reacted to a loss of altitude by pulling on the stick, which would have been an appropriate reaction with the autopilot fully enabled, which would then have put the aircraft in a climbing configuration. However, the airplane's systems had entered a mode of lesser automation ("direct law" in Airbus terms) due to a blocked airspeed sensor, allowing the pilots to put the plane in a nose-high stall configuration, from which they did not recover.
- On January 17, 2015, the offshore supply vessel "Red7 Alliance" collided with a lock gate of the Kiel Canal in Germany, damaging it severely. An investigation concluded that the levers controlling the ship's Azimuth thrusters were not used in a way appropriate to the mode they were set to, resulting in the ship accelerating instead of coming to a stop in the lock.
- On August 21, 2017, the US Navy destroyer USS John S. McCain collided with a commercial tanker in the Strait of Malacca, resulting in the loss of life of ten crew members. An investigation conducted by the US military concluded that immediately prior to the collision, helm and propulsion controls had been redistributed between bridge stations, and the bridge crew was not fully aware of that redistribution.
- On April 10, 2018, the 5000 ton supply vessel VOS Stone unberthed from a wind platform under construction in the Baltic Sea. The vessel's master decided to put the steering in an alternative mode to perform a test of the system. Insufficient communication with the officer of the watch led to a temporary loss of control, collision with the platform, injury to three crew members, and significant damage.
Modes are intended to grab the user's full attention and to cause them to acknowledge the content present in them, in particular when critical confirmation from the user is required. This later use is criticised as ineffective for its intended use (protection against errors in destructive actions) due to habituation. Actually making the action reversible (providing an "undo" option) is recommended instead. Though modes can be successful in particular usages to restrict dangerous or undesired operations, especially when the mode is actively maintained by a user as a quasimode.
Modal proponents may argue that many common activities are modal and users adapt to them. An example of modal interaction is that of driving motor vehicles. A driver may be surprised when pressing the acceleration pedal does not accelerate the vehicle in the forward direction, most likely because the vehicle has been placed in an operating mode like park, neutral, or reverse. Modal interfaces require training and experience to avoid mode errors like these.
Interface expert Jef Raskin came out strongly against modes, writing, "Modes are a significant source of errors, confusion, unnecessary restrictions, and complexity in interfaces." Later he notes, " 'It is no accident that swearing is denoted by #&%!#$&,' writes my colleague, Dr. James Winter; it is 'what a typewriter used to do when you typed numbers when the Caps Lock was engaged'." Raskin dedicated his book The Humane Interface to describe the principles of a modeless interface for computers. Those principles were implemented in the Canon Cat and Archy systems.
Some interface designers have recently taken steps to make modal windows more obvious and user friendly by darkening the background behind the window or allowing any mouse click outside of the modal window to force the window to close – a design called a Lightbox – thus alleviating the risk of modal errors. Jakob Nielsen states as an advantage of modal dialogs that it improves user awareness. "When something does need fixing, it's better to make sure that the user knows about it." For this goal, the Lightbox design provides strong visual contrast of the dialog over the rest of the visuals. However, while such a method may reduce the risk of inadvertent wrong interactions, it does not solve the problem that the modal window blocks use of the application's normal features and so prevents the user from taking any action to fix the difficulty, or even from scrolling the screen to bring into view information which they need to correctly choose from the options the modal window presents, and it does nothing to alleviate the user's frustration at having blundered into a dead end from which they cannot escape without some more or less destructive consequence.
Larry Tesler, of Xerox PARC and Apple Computer, disliked modes sufficiently to get a personalized license plate for his car that reads: "NO MODES". He has used this plate from the early 1980s to the present, on various cars. Along with others, he has also been using the phrase "Don't Mode Me In" for years as a rally cry to eliminate or reduce modes.
Bruce Wyman, the designer of a multi-touch table for a Denver Art Museum art exhibition argues that interfaces for several simultaneous users must be modeless, in order to avoid bringing any single user into focus.
Avoid when possible
Alternatives to modes such as the undo command and the recycle bin are recommended when possible. HCI researcher Donald Norman argues that the best way to avoid mode errors, in addition to clear indications of state, is helping the users to construct an accurate mental model of the system which will allow them to predict the mode accurately.
This is demonstrated, for example, by some stop signs at road intersections. A driver may be conditioned by a four-way stop sign near his or her home to assume that similar intersections will also be four way stops. If it happens to be only two way, the driver could proceed through if he or she sees no other cars. Especially if there is an obstructed view, a car could come though and hit the first car broadside. An improved design alleviates the problem by including a small diagram showing which of the directions have a stop sign and which don't, thus improving the situational awareness of drivers.
Modal controls are best placed where the focus is in the task flow. For example, a modal window can be placed next to the graphical control element that triggers its activation. Modal controls can be disruptive, so efforts should be made to reduce their capacity to block user work. After completing the task for which the mode was activated, or after a cancel action such as the Escape key, returning to the previous state when a mode is dismissed will reduce the negative impact.
In the book The Humane Interface, Jef Raskin championed what he termed quasimodes, which are modes that are kept in place only through some constant action on the part of the user; such modes are also called spring-loaded modes. The term quasimode is a composite of the Latin prefix quasi- (which means almost, to some degree) and the English word "mode".
The application enters into that mode as long as the user is performing a conscious action, like pressing a key and keeping it pressed while invoking a command. If the sustaining action is stopped without executing a command, the application returns to a neutral status.
The purported benefit of this technique is that the user doesn't have to remember the current state of the application when invoking a command: the same action will always produce the same perceived result. An interface that uses quasimodes only and has no full modes is still modeless according to Raskin's definition.
The StickyKeys feature turns a quasimode into a mode by serializing keystrokes of modifier keys with normal keys, so that they don't have to be pressed simultaneously. In this case the increased possibility of a mode error is largely compensated for by the improved accessibility for users with physical disabilities.
- Usability Glossary: modeless Archived 2007-10-22 at the Wayback Machine
- Usability Glossary: mode error
- "How to Use Modality in Dialogs". Oracle Corporation.
- Glossary: mode error
- Usability Glossary: modal
- Poor UI design can kill
- BEA final report on the loss of Air France 447
- M/V Red7 Alliance investigation report (German)
- "USS McCain collision ultimately caused by UI confusion". 2017.
- Investigation Report 118/18 (pdf), Federal Bureau of Maritime Casualty Investigation (Germany), April 10, 2019
- Infragistics.com. "Modal Panel - Context". Archived from the original on 2013-05-06.
- Aza Raskin, A List Apart: Never Use a Warning When you Mean Undo
- Jakob Nielsen, Alertbox. "10 Best Application UIs".
- Origins of the Apple Human Interface by Larry Tesler, Chris Espinosa
- Technology for Experience's Sake: Guest Post by Bruce Wyman
- Bruce Wyman's post at the ixda.org mailing list
- Infragistics.com. "Modal Panel - Implementation". Archived from the original on 2013-05-06.
- "Design rules based on analyses of human error".
- Usability Glossary: spring-loaded mode
- Spring-Loaded Modes, Jakob Nielsen.
- Buxton, William A. S. (1995). "Chunking and phrasing and the design of human-computer dialogues". In Baecker, Ronald M.; Grudin, Jonathan; Buxton, William A. S.; Greenberg, Saul (eds.). Readings in human-computer interaction : toward the year 2000 (2 ed.). San Francisco, Calif.: Morgan Kaufmann. pp. 494–499. ISBN 978-1-55860-246-5. acmid 212970.
- Modelessness in UsabilityFirst glossary
- Modelessness in Apple's HIG guidelines
- Definition of mode error at Usability First
- An Example of a mode error in Excel
- John Rushby. Using Model Checking to Help Discover Mode Confusions and Other Automation Surprises. A paper discussing an automatic method for locating mode errors.
- Jakob Nielsen on Spring-loaded modes