On Sat, 9 Sep 1995, Andrew Davidoff wrote: > On Sat, 9 Sep 1995, Paul Cole wrote: > > Interesting idea, autocompletion. However, wouldn't it be much easier to > > deal with autocompletion by assuming that anywhere a noun could go, if it > > doesn't match after they hit return <i.e. during normal command > > processing> to try to complete it, and if not unique, send some message > > back to the user stateing as much, maybe even a list of the possible > > completions available given the partial input. That way you don't have > > to fool around with character at a time input. > > actually, i think the guy was using [tab] as it's used in, for instance, > zsh if yer familiar with it. you hit tab in the middle of a file-name and > you can set it to show you a menu of auto-completion-choices and cycle > through them... anyway, i think he was using as a kinda substitute or > add-on to abbreviations. what yer talkin about _is_ abbreviations, and > they _are_ simple :) this guy's idea entailed never executing the > command at all, until auto-completion (or really just completion) had > dealt with it if necessary (if they hit tab). > -dert Well, the problem with using tab is, your going to have to use character at a time input which will complicate your input routines significantly and won't be acceptable at all for those who use clients as clients like tinyfugue do not send data to the remote until a newline is read on the on the local system. So you won't see a tab until you get the entire line of data and on tinyfugue at least, it doesn't even bother to send tabs at all. Btw, tcsh and bash also support completions of various sorts. --Paul
This archive was generated by hypermail 2b30 : 12/07/00 PST