In article <x7fz2rlw4q.fsf@bonnet.wsrcc.com>,
Wolfgang S. Rupprecht <wolfgang+gnus20041129T213431@dailyplanet.dontspam.wsrcc.com> wrote:
>
>atatat@atatdot.net (Andrew Brown) writes:
>> *lots* of programs use $PATH and and $EDITOR (or $VISUAL -- why are
>> there two?)
>
>Thats an interesting question.  At one time back when ~1-MIPS
>VAX-750's roamed the land sucking down 10kw a piece and burping out
>enough heat to make ones home oven blush, things weren't quite as
>snappy as they are now.  Real (tm) full-screen editors took many
>seconds to start up.  Emacs took 10 to 15 seconds of CPU time to
>start.  VI wasn't all that much better.  One devised all sort of
>crocks to keep from starting a second copy of *the* editor (see
>emacsclient).  It just wasn't considered sporting to use that much CPU
>time for trivial editing tasks.  The env variable VISUAL was for
>"important" things that would justify a large CPU expenditure.  The
>variable EDITOR was for simple things that could be done by much
>simpler/faster programs.
>

Actually it was not an issue of CPU speed. Back then the only game
in town was vi; emacs came later. That is why the name "VIsual".
Vi loaded sportingly fast too because it was small and had the
sticky bit set. It was more of an issue if your terminal could
handle it, or if your connection to the machine was tolerable. It
was not a lot of fun trying to startup vi on a decwriter III, or
through a 300 baud modem. For the second case vi had the "slowopen"
hack to show fewer lines than the terminal height; for the first,
you could use ex. They were also people around who were ed freaks.
One of them wrote a whole book in ed, and considered vi to be a
CPU waste.

christos