User:Natasa Siencnik/notes/gentner/

Abstract

Don Gentner / Jakob Nielson: The Anti-Mac Interface. In: Communications of the ACM. August 1996, Vol. 39, N°8.

Introduction

1. Alternative Interfaces
   • human interface is stuck, very little innovation in interface design anymore
   • user has settled on the WIMP (windows, icons, menus, pointer) model
   • violating basic assumptions as useful mental exercise to find new concepts
   • article explores possible types of interfaces that could result in such a violation
   • focus on Macintosh interface as prime example of current interface paradigm
2. Macintosh Design
   • designed for naive users without any previous computer experience
   • targeted at a narrow range of applications (office work, entertainment, multimedia)
   • weak computational resources (computer with 128KB RAM, 400KB storage device, printer)
   • supported by poor communication channels (screen, keyboard, mouse)

Human Interface Design Principles

• based on fundamental principles of human-computer interaction
• how do these principles limit the computer-human interface?

1. Metaphors from familiar non-computer world
   • computer files represented as documents in paper folders placed on desktop
   • files deleted by dragging them to the trash can
   • trying to overcome the limitations of the desktop metaphor (rooms, buildings, village)
   • automobiles have developed their own interfaces without metaphors based on horses
2. Problems with Metaphors
   • target domain has features not in the source domain (e.g. replace command in text editor)
   • source domain has features not in the target domain (e.g. marking of typewriters)
   • some features exist in both domains but work differently (e.g. white space as character)
3. Trash Can Metaphor
   • single-trash-can metaphor has led to a system that fails to meet user's needs
   • to avoid the confusion of multiple trash cans in the interface, all trash cans are combined
   • when user empties the trash to create room on a floppy disk, both trash cans are deleted
   • this is a limitation that does not serve the user's real needs
4. Desktop Metaphor
   • save training time by taking advantage of having learned to operate a traditional office
   • next generation of users will make their learning investments with computers
5. Direct Manipulation
   • users interact directly with objects in the interface (e.g. drag and drop a document)
   • user cannot group a related series of basic actions into one high-level action
6. See-and-Point
   • users interact with computer by pointing at the objects they see on the screen
   • we have lost all power of language, and cannot talk about objects that are not visible (yet)
   • utilize more of the power of language to communicate more precisely
7. Consistency
   • difficult to apply in real situation with wide array of conflicting things
   • learning will be reduced if objects with similar function always look / behave alike
   • in real life we have a wide range of appearances and they can still be easily recognizable
8. WYSIWYG
   • what you see is what you get > document on the screen will look the same when printed
   • problem that it is usually equivalent to WYSIATI (what you see is all there is)
   • WYSIWYG document only shows final printed representation, not the user's intentions
   • Standard Generalized Markup Language (SGML) preserves semantic meaning
   • might be useful to have a different representation when preparing the document
   • electronic information should be modularized and enabled to link to backup information
9. User Control
   • many situations where we do not want to be in control > delegate to machines
   • even if control is desirable, full control is becoming impossible with networked computers
10. Feedback and Dialog
    • computer interface should provide the user with clear and immediate feedback on actions
    • delegation of sequence of activities to an agent or encapsulated in a script
    • then there is no longer a need for detailed and continuous feedback
11. Forgiveness
    • user actions should generally be reversible and users should be warned of data loss
    • computer needs to build a deeper model of our intentions and history
12. Perceived Stability
    • elements in the computer interface should not be changed without user's involvement
    • but dividing control between the user and the computer or networked computers
    • stability can be boring
13. Aesthetic Integrity
    • graphic design of the interface should be simple, clean, and consistent
    • if computers could communicate with richer language, look would be less important
    • richness can increase usability by making it easier to distinguish
14. Modelessness
    • computer interface should not have distinct modes that restrict the user's actions
    • users should be able to perform any task at any time
    • problem of modelessness > user cannot cope with everything at once

Anti-Mac Interface

1. Central Role of Language
   • language lets us refer to things not immediately present, potential actions
   • encapsulate complex groups of objects or actions and refer to them with single name
   • ability to deal with imprecise language will increase the computer's flexibility
   • command line interfaces have some of the advantages of language
   • command line interfaces have rich syntactic structure to form complex commands
   • but only a limited number of commands and cannot tolerate synonyms, misspellings
2. Richer Internal Representation of Objects
   • current interfaces have access to very little information about the used objects
   • the only information about a file may be its name, size, and modification date,
   • the type of data it contains, and the application that created it
   • but it could include its authors, topic matter, keyword, importance, relations
3. More Expressive Interface
   • richer internal representation of objects will allow more intelligent interpretation
   • but it will also be reflected in a more expressive interface with external representation
   • variety adds visual interest and helps us quickly to locate e.g. books in a shelve
4. Expert Users
   • think about trade-offs between ease of learning and power in new interfaces
5. Shared Control
   • computer-based agents have not yet progressed beyond simple tasks
   • users still need to determine the balance of control in their environment

Conclusions

• Macintosh interface as a starting point for considering future needs of user interfaces
• outlines are impractical with the current state of computers and software
• language understanding is still in its infancy
• interface designers need to start work for the next generation of computers