Peter,
By now I guess you are getting the idea that we all consider this to be
a "difficult" problem that you have before you, and IT IS... with no 'magic
bullet'. Nonetheless difficulty is relative:
If the budget is infinite, no problem!
If the time is infinite, no problem!
If the quality required is zero, no problem.
Assuming that the situation in your case is the opposite of the above, you have
just one more card to play: wisdom!
A wise choice of human, hardware and software resources, combined with luck can
accomplish near anything, luck being an important element!
The following 'wisdoms' I consider to be unavoidable in your case:
1. Find someone who's done near the same thing before. Spend $. Estimate.
2. Prototype very early, and let the prototype grow into the solution.
3. Obtain the 'buy-in' of the ultimate client... your boss' boss?
4. Use automation. The prompts that I referred to should be structured.
5. Set aside all else, and focus on one (elaborated) goal.
6. Use Jess and Java... keep each scenario as simple as possible.
7. Get most of the books mentioned by James Owen, but keep it practical.
8. Use UML for your object design, with no more than 9 symbols on a diagram.
9. Make sure that you do backup, and that ALL project files zip up as one!
David
Quoting [EMAIL PROTECTED]:
> You must narrow the possible options down by defining the data
> structures, key concepts, and general conventions to be used by the
> specific system, before you can hope to produce this simple
> description. Otherwise you're proposing to write an entire new manual
> for Jess, because you're leaving things wide open. The Jess language
> is a reasonable sized and complete programming language, and
> describing everything it can do basically means teaching the whole of
> that language. The only way to simplify it is to choose a subset of
> the possiblities, and doing that is going to take some knowledge
> engineering.
>
>
> I think Peter Frederick wrote:
> [Charset iso-8859-1 unsupported, filtering to ASCII...]
> > Dear David,
> > The big problem I have is I haven't got a clue about the needs of the
> rules
> > base. We are using JESS to develop a way of creating and negotiating
> > process control systems for factories and want to capture the knowledge
> of
> > the people who design the system. The number of rules will be immense
> and
> > so it is unreasonable just to get natural language replies back from the
> > experts as it will take too long to code them up. What I need is a
> simple,
> > well structured description of how to create a knowledge-base entry and
> how
> > to create a rule (giving all the possible options). This is what I am
> > trying to compile from the manual but it is proving difficult.
> >
> > If you have any further thoughts/links/information, I would be very
> grateful
> > to hear from you.
> >
> > Yours,
> >
> > Pete Frederick
> >
>
>
>
> ---------------------------------------------------------
> Ernest Friedman-Hill
> Distributed Systems Research Phone: (925) 294-2154
> Sandia National Labs FAX: (925) 294-2234
> Org. 8920, MS 9012 [EMAIL PROTECTED]
> PO Box 969 http://herzberg.ca.sandia.gov
> Livermore, CA 94550
> ---------------------------------------------------------------------
> To unsubscribe, send the words 'unsubscribe jess-users [EMAIL PROTECTED]'
> in the BODY of a message to [EMAIL PROTECTED], NOT to the
> list (use your own address!) List problems? Notify
> [EMAIL PROTECTED]
> ---------------------------------------------------------------------
>
---------------------------------------------------------------------
To unsubscribe, send the words 'unsubscribe jess-users [EMAIL PROTECTED]'
in the BODY of a message to [EMAIL PROTECTED], NOT to the
list (use your own address!) List problems? Notify [EMAIL PROTECTED]
---------------------------------------------------------------------
