SSteinberg.SoftArts%MIT-MULTICS@sri-unix.UUCP (07/15/83)
A - Icons are here to stay. Right now icons are used to provide a clear method of interacting with stored data objects just as multiple windows provide access to parallel processes. I remember people looking at EMACS in the early seventies and crabbing that screen editors wasted system resources and that they were just a fad. Screen editors were not a fad since they made things easier to work with and visualize. The difference between a screen editor and a "line editor" is the difference between a map and a program that will answer questions about cartographic relationships. [I recently read a paper from Carnegie Mellon (1981?) in which they compared expert and non-expert use of an editor and found that the number of keystrokes to complete a task was about the same but that expert users used more sophisticated commands to cut the amount of output they had to look at and so sped up their interaction. My guess is that icon based systems will cut the difference between advanced and naive users.] B - Yes, some are too cute. When MIT first got its XeroGraphic Printer everyone had to print everything in 70-80 different fonts (excuse my exageration). Some of them were absolutely awful and the mixture of Western, Gothic, Helvetica and whatnot was a bit much. Things have settled down a fair bit but multiple fonts are still in use. C - Having studied the maturation processes of insects, cheeses, wines, animals, people and scientific theories I would guess that what must happen for this field to mature is the passage of time, meaning time spent using these things. This means that these systems have to get out into the field and get tested in vivo. [Steven Gould recently had a couple of papers on diversity and maturation, one published in Vanity Fair and the other in Natural History. The first asked to figure out where all the .400 hitters (in baseball) went and the latter why so many old fossils don't fit into modern phyla. In a new field a thousand flowers bloom hence .200 and .400 batters or animals with basically different designs. As the field matures and the game progresses the survivors learn (excuse the teleology) what it takes. You can't hit to the empty spot in left field if the left fielder has figured out where he should stand. Gould also points out that the "standard" or current solution need not be optimal and as a matter of fact often isn't.] D - Premature standardization could murder the field. Unlike the audio cassette business I think standardizing icons now would be premature. I am sure that this will happen in the future though there is no point standardizing everything, just the important ones. The basic driving mechanisms in automobiles are pretty much standardized but it will probably take another 1000 years before bumper heights are matched. [There was an article on annealing as a means of attacking the traveling salesman problem in which they point out that too rapid cooling results in bad solutions.] E - I'll argue that it is obvious when to use icons as opposed to words but we don't know what they are yet. Einstein's theory of relativity is obvious once it is explained carefully. As a bird watcher I am quite familiar with the obvious. F - When designing an interface it is necessary to consider the objects which are to be presented to the user. Icons are little tokens which behave like checkers or playing cards. This makes them a very useful way of presenting files in a directory which share many token-like properties (identity and manipulability for example). Unfortunately a lot of people are going to fall in love with icons and try to make them do everything. I play bridge with a deck of cards but I eat with a fork.