And to make things more confusing, interface metaphors have multiple purposes. Not only are interface metaphors a guide through a new system, but they also help mask an abstract concept with a "real world" object.

Interface metaphors work as user introduction to an unfamiliar system. They are a shortcut, and method of flattening the learning curve for a new user to USE the system. Interface metaphors use terms from one system, and apply them to another, and by doing so, hasten usablity.

Interface metaphors have another seperate function, which is to mask complex processes in simple, concise terms, thereby controlling complexity. Something as terse as a pipe [ | ] in UNIX obscures the fact that the operating system is performing a complicated series of commands that spawns Input/Output redirection. Heck, most times we don't want to know exactly what is happening behind the metaphor.

Each of these purposes comes with its own set of problems, its own set of questions.

What happens when, after becoming proficient with a system, the metaphors that were used to learn the system start to fall apart?

What happens when the system in which the metaphor was modelled after becomes obsolete? Let's say for the sake of argument that human being no longer used trash cans, but the trash can on the PC desktop was still used? The metaphor loses it's meaning.

With computer systems getting more and more complex, there are layers and layers within layers of metaphors. Is it any wonder this gets confusing?