Re: LFM + LFSP = LFE?

["Neil W. Van Dyke" <address@hidden>]

Geoffrey Knauth <geoff@knauth.org> writes at 10:38 12-Jun-2003 -0400:
> When I think about it, newcomers to Emacs too (and many other pieces of 
> software) are fearful of the damage they may inflict, until they are 
> taught how to fix mistakes, or return to a previous known state, at 
> which point they cast away fear and launch full speed ahead with 
> experimentation.

Good point.  I see this perceived fragility of Emacs and Smalltalk
environments as an artifact of their high hackability factor.  Jumping
in and making enhancements to the running system, learning as you go, is
easy.  And programmers love to program programs for programming.

This degree of hackability is one thing I've found missing from one of
my current tools, DrScheme.  They've done such a good job of
compartmentalizing the Scheme environments, and imposing modularity
abstractions, that interactively hacking the environment itself seems
hard.  (DrScheme does have some extension facilities, but doesn't
encourage you to play around with arbitrary system code from within the
running system.)  They have good reasons for making DrScheme an
unbreakable tool, of course, but I'd like to combine the goodness of
DrScheme with the hackability of Smalltalk-80, Emacs, Self, Lisp
machines (?), etc.  Perhaps atop Matthew Flatt's continuing
Scheme OS work.  I suspect this is as much an HCI problem as a PL one.

-- 
                                             http://www.neilvandyke.org/