[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Multiple names for "the same" file



    Date: 11 Apr 1982 1943-EST
    From: Robert W. Kerns <RWK at SCRC-TENEX>
    Subject: Re: Multiple names for "the same" file
    To: MOON at SCRC-TENEX, AGRE at MIT-AI
    cc: BUG-ZWEI at MIT-AI, RWK at SCRC-TENEX
    In-Reply-To: Your message of 11-Apr-82 1426-EST

    How about a warning (in the echo area, without beep) if you read in a file
    when you have another of the same NAME and TYPE but different VERSION?
    Also, if two files of the same name, one type :UNSPECIFIC and one type
    anything else.  The check should also occur before renaming a file with
    ^X^W, and more actively complain if the filename would in fact be duplcated,
    since this is almost certain to leave you totally confused.

     Not only that, but same name/etc but different FILE SYSTEM too!  So it
should warn you on AI:AGRE;FCFLOW > and FC:AGRE;FCFLOW.#> even if they have
different version numbers, in case you are keeping duplicate copies on different
machines.  You should never have two files on different machines with the same
name (modulo filename syntax) and different contents; doing so is very likely to
cause confusion.  - phIl
     P.S.  How about different file system -plus- different directory?!  How far
do you go?  Maybe it should offer to source-compare in each of these cases, not
as a question you have to answer NO to but something like "if you want a
source-compare in a buffer, type super-c" or some such.  That'd be really nice
sometimes.
     P.P.S.  In general there should be a theory about when the machine should
make a point of advertizing one of its features to you.  Already it's rather
hard for a semi-frequent user of the LispMachine to keep track of all the
commands and functions s/he needs to know about to reasonably comfortably get
through the day.  Maybe there already is such a theory?  Certainly there is a
well-developed theory of notifications (even if I grouch at it at places).