Actions for Multiple Objects

Examples:

Moving a set of graphic objects on a WYSIWYG editor
Copying a batch of files from one place to another
Locking all four car doors from the driver's seat

Context: The artifact contains multiple real or virtual objects, such as files, or CDs, or car windows. There are actions to be performed on those objects, and users are likely to want to perform actions on two or more objects at one time.

Problem: How can the artifact make repetitive tasks easier for the user?

Forces:

People aren't good at repetitive tasks; machines are.
Users don't want to spend time performing the same action to every single object separately.
It might also be error-prone to do this, if the action is complex or requires precision.
Different objects may have different sets of actions.
Putting together a collection of objects may take time and mental effort.

Solution: Allow the action to be performed "in parallel" across a set of user-selected objects. Make it easy to put that collection together. Multiple-selection is the typical way to do it today in computer-based artifacts: you choose objects by shift-clicking on them or lassoing them, for instance, and then you perform the action (Localized Object Actions). You could also provide a separate facility for performing an action on all instances of a certain kind of object, such as a "Clear All" button on a drawing program.

What if the user has a set of multiple objects that don't share the same actions? You could disable the actions that aren't relevant to all of them (Disabled Irrelevant Things), but the user may not understand why they're disabled. A crystal-clear model of the object types and their available actions would help the user avoid this situation in the first place. Start there.

Resulting Context: If the action is difficult or complex, you might want to treat the set of objects to be acted upon as a form of Editable Collection; but it would be overkill for really simple actions. In short, don't make the user spend more effort putting together the collection of multiple objects than they would have spent performing the action on each individual object!

Comments to: jtidwell@alum.mit.edu
Last modified May 17, 1999