[comp.ai] expert system building tools

kk@aiva.ed.ac.uk (Kathleen King) (05/20/88)

I'm trying to find out what folk think of various expert system building
tools they have experience with. If there's enough interest I'll post
the results to the net. Ta.

Do you now or have you ever used any of the following tools?

Acquaint (A.K.A 'Daisy')
APES
Arity Expert System Development Package
ART
Auto-Intelligence
Crystal
DUCK
ENVISAGE and SAGE
ES Environment
ESP advisor
ESP Frame Engine
EST(Expert Systems Toolkit)
Experkit
ExperOps
Expert Controller
Expert Ease/Super Expert
Expert Edge
Exsys
Ex-Tran 7
1st Class
1st Class Fusion
Flops
GEST (Generic Expert System Tool)
GOLDWORKS
GURU
G2
HUMBLE
Insight 2+
Intelligence/Compiler
KDS 3
KEATS (Knowledge Engineer's Assistant)
KEE
KES (Knowledge Engineering System)
Keystone
Knowledge Craft
Knowledge Workbench
Knowol
Leonardo
Lisp In-Ate/Micro In-Ate
LOOPS
M1
MacKIT
MicroExpert
Muse
Nexpert/Nexpert Object
Nexus
OPS5
OPS83
Personal Consultant Easy
Personal Consultant Plus
PICON
RuleMaster 2
Savoir
Super Expert
S1
TIMM
TOPSI
TWAICE
VP Expert
Wisdom XS
Xi Plus
XPER
XSYS


If so I'd greatly appreciate hearing your answers to the following questions.

1) How long did it take to learn?
2) Did you teach yourself or get 'learning support'?
3) Do you still use it?
4) Would you choose it again or something else?
5) What sort of application have you used it for?
6) Have you used it for more than one application?

I realise that many PC users who might have these tools do not have access to 
the net. Second hand information from them is just as valuable.
Thanks verrrry much.

noren@dinl.uucp (Charles Noren) (09/13/88)

Dear Kathlenn King,

I'm new to usenet, so I just saw your article posted May 20, 1988
about wanting information about tools. (I hope this reaches you)

I've been working with G2 by Gensym since March-88.  While I have
*never* used expert shells before, we did evaluate it against ART,
KEE, and Savoir.  We choose G2 because:

  1. We were not strong LISPers;
  2. G2 seemed easy to learn;
  3. G2 provided good graphics;
  4. Gensym seemed to tailor G2 towards control of real-time devices.

G2 has lived up to our expectations.  I took the 1 week G2 course in
Cambridge MA and found it easy to learn.  I took the course in Jan-88
and it was not until March-88 when I first used it.  I had to brush up
a little on G2, but the documentation and thier examples helped quite
a bit. G2 is easy to use (which is something since I am a neophyte to
this technology).  Its rulebase language has a natural langauge syntax.
G2 uses a mouse driven context sensititive editor to prompt you for
everything in it.  I have never felt that the editor got in the way as
I better understood the system.

Our primary application for G2 is network management prototyping for
a research project I'm on (we are using it on a Sun 3 workstation).
G2 works very well for this application.
There are a few things that would be nice if they were added, but I
find that I can work around it quit well.  Gensym does not allow
modifications to G2 from its underlying LISP environment as ART or KEE
does (because Gensym is writing "garbageless" code).  We do not mind
since all of us in the group are C hackers.

Currently we are waiting for the next release of G2 that will permit
our C device servers to interface with G2.  This is due (I am told
by Bob Moore, the president of Gensym) around the end of Sept-88.

I am very pleased with G2 and the support I get from Gensym.  I will be
reporting my experiences with G2 at Milcom-88 in San Diego, CA at the
end of Oct-88.  We do plan on using G2 for other applications prototyping.

I hope this helps --

    Chuck Noren
    Martin Marietta I&CS
    Denver, CO
    (303) 971-7930

(PS.  I would like to find out what information you got on the topic).
-- 
Chuck Noren     (303) 971-7930