OverviewHuman Computer InteractionDesign Principles

Shneiderman, Ben (1983)

Direct Manipulation. A Step Beyond Programming Languages

IEEE Transactions on Computers, Vol. 16, No. 8, August, pp. 57–69

Google this publication · ScholarGoogle this publication

Related Topics: Interactive Learning Environments

Review by: Dreier, Matthias (2005-02-02)

In 1983 most computer programmes were operated through a special command language. Only a few programmes supported the use of a mouse or a joystick, the most successful ones were video games. In this seminal article Ben Shneiderman outlines the potential of direct manipulation for a wide range of applications. A reprint of the article can be found in Noah Wardrip-Fruin and Nick Montfort (2003): The New Media Reader, MIT Press.

Shneiderman begins with two prime examples of office applications: word processing and spreadsheet. Direct manipulation in this context means being able to display the document in its final form, to move a cursor by arrow keys, to see the results of an action immediately, and to easily reverse commands. Especially novices enjoy using direct manipulation software because they rapidly get a feeling of mastery of the system. The author continues with further examples of direct manipulation software. The most promising example is what we now call the “desktop metaphor,” first seen in Xerox’ Star and later in Apple’s Lisa and Macintosh computer.

According to Shneiderman direct manipulation works better than command line operation because it helps the user to deal with the problem directly. The suitable representation is crucial to problem-solving and learning. An early example of visual learning software is Papert’s Logo Language (turtle graphics). Another example is Stanley Smith’s CDC Plato, a simulation of a chemistry laboratory.

There are also problems with direct manipulation: wrong visual information can lead to great confusion, icons require more screen space than textual descriptions, and users might draw incorrect conclusions from simplified analogies. Shneiderman proposes ample user testing before introducing a direct manipulation system, especially if uncommon metaphors and analogies are used.

Shneiderman introduces the Syntactic/Semantic Model as a possible explanation for the users’ mode of operation and the way direct manipulation works. Novices mostly think in terms of syntactic commands, experts have a hierarchically structured knowledge of semantic concepts. Syntactic here denotes a series of memorised commands whereas semantic denotes the concept of a function, e. g. “move cursor, cut, move cursor, paste” vs. “rearrange the words of a sentence.”

Shneiderman’s article on direct manipulation is a milestone in human computer interface design and blazed the trail towards modern user interfaces using mouse and visual representations. This article brilliantly illustrates the enthusiasm at that time. Nowadays the reader might not be able to grasp that enthusiasm because the paradigm of interaction has almost completely shifted from command line interface to direct manipulation. However not all human-computer interaction today is direct manipulation. For instance, many applications now feature a macro language and software is still largely written as lines of codes rather than visually programmed. Therefore Shneiderman’s critical remarks on the limits of direct manipulation are still valuable today.