jans@tekgvs.GVS.TEK.COM (Jan Steinman) (11/02/88)
(Posted for a friend -- please reply to <cliffm@tekgvs.GVS.TEK.COM>.) For a Software Productivity project, how about standardizing on EXIT terminology. Each user has to blast away at the keyboard and try all this 'QUIT' stuff unless he already knows how to end a program. Where is standardization? Does everyone writing software need to exercise individuality, help avoid copyright suits, or please their company's idiosyncrasies? 'QUIT' LIST usage: E, EXIT many ESC Generic CTRL-C Basic Programs and many others CTRL-D Logouts CTRL-Z File end marker & exit Q, QUIT many DONE Logouts BYE UNIX SYSTEM BASIC F1 Typical of new IBM PC software F10 same CTRL-] Kermit, Sytek MOUSE using graphics boxes, with click LOGOUT UNIX LOGOFF Mainframes CTRL-K X Wordstar, SPRINT, others X same abort? ? Abandon Mech CAD, Orcad KILL -9 # UNIX lprm -Plw# UNIX WINDOWS New UNIX, PC ICONS Mac, others Suspend Orcad RESET Most minis Terminate MSDOS Batch TIMEOUT Final remedy #1 CTRL-ALT-DELETE Final Remedy #2, IBM PC's POWER SWITCH Final Remedy #3, common to most systems --and others. List those you know. The auto industry has reasonably standardized clutch, brake, and ignition key locations and direction of push or rotation. It took a while. How did they do it? This 'QUIT' list needs to be cut by a factor of 2 or 3 immediately, and by industry agreement to preferred terms. New software now needs to include two exit paths, one the local idiosyncrasy path plus an industry standard term for exit of any program. The user can then take his choice and always "get there". PS: Reply to your favorite standards committee or working group. :::::: Software Productivity Technologies -- Experiment Manager Project :::::: :::::: Jan Steinman N7JDB Box 500, MS 50-383 (w)503/627-5881 :::::: :::::: jans@tekcrl.TEK.COM Beaverton, OR 97077 (h)503/657-7703 ::::::
barmar@think.COM (Barry Margolin) (11/03/88)
I saw many problems and misconceptions in the "quit list" that was
posted. One misconception was to label "ctl-D" as performing logout,
presumably referring to Unix. ctl-D does NOT mean logout in Unix, it
means end-of-file; most Unix programs exit when they reach the end of
their input stream, and when a login shell exits you are logged out.
A general problem with the list was that many different concepts were
being lumped under the name "quitting". For instance, the list
included Kermit's ctl-] operation, which is merely an escape that
suspends the terminal emulation portion of the program so that the
user can issue Kermit commands. In order to quit the program the user
must still type "quit".
I agree with the general philosophy that this is a simple area that
could, and probably should, be standard. However, different user
interface styles frequently force differences even in such a simple
thing. For instance, if a program author wants to provide one-letter
abbreviations for common commands, and Q is being used for something
already, he might choose to call his quitting command "Exit" so that
the "E" abbreviation can be used. User interface issues are too
complicated to try to standardize on anything at this point.
Jan included an analogy with cars, pointing out that all cars have the
gas, brake, and clutch pedals in the same places. I suspect this
wasn't true in the first decade or so of the automotive industry, and
computer user interfaces are still at that stage (mostly because
computers keep changing so quickly that it not useful to try to
standardize). I also think that car pedals are probably more properly
compared with computer keyboards, which are much more standardized.
To illustrate that the automotive industry still has user interface
problems, I got into my mother's car a couple of weeks ago and spent
fifteen minutes finding all the light controls.
Barry Margolin
Thinking Machines Corp.
barmar@think.com
{uunet,harvard}!think!barmar