Object Oriented CASE Tools

[<@IU.AI.SRI.COM:slug-admin@iu.ai.sri.com>]

    Date: Tue, 26 Jan 1993 11:22 CST
    From: Donald H. Mitchell <dmitchell@amoco.com>

	.... in all cases, although the
    tools are good, they lack active participation (e.g., gen'g
    defuns and defmethods from documentation specs; having graphical
    class/method browsers). 

Several years ago, soon after the first incarnation of the Flavor
Examiner (which didn't have any display of inheritance hierarchy), 
I wrote a flavors grapher as an extension.  It allowed the user to
graphically browse inheritance hierarchies, and associated all the
standard FlavEx commands with the nodes via popup menus.

Subsequently, Tom Shepard took the code, added some interactive node
movement and ancillary display capabilities, and implemented a "High
Order Software" graphical development tool.  This allowed the user to
graphically build a program structure (call tree), specify arguments,
etc., in a system that was intended to generate code.  I don't know 
what the fate of this system was or whether it ever actually produced
code, but the formalism that it was made for was intended specifically
for that purpose.

Later Genera releases and substantial changes to the Flavor Examiner
made my code obsolete, although someone out there may have maintained a
working version.  I could resurrect and update the code (I may even have
a copy of Shepard's somewhere) if I were given an adequate incentive.  I
always intended to make a flavors (CLOS) development package out of it,
but other priorities took over.

Of course, the right thing now would be quite different from what it
was then.  I have a newer, better grapher which is readily adaptable for
such purposes, but again I have other priorities which tend to keep me
from pursuing this independently.  However, it strikes me that visual
programming environments are in a sense the ultimate in CASE tools, and
that it would be a lot of fun to work on such a beast.  All I really
need is some funding (hint, hint).