|Introduction to Expert System Shells|
|Expert System Shells|
The theory of expert systems could not sound better. Derived from work in the science of artificial intelligence they are computer programs that are designed to mimic the working of a human expert. And, in theory, that should mean that, whenever you want access to expertise in some particular field, you can simply turn to your computer, switch it on, and have instant expertise there on the screen before you. But does it always work like that?
Consider an example:
A skilled medical practitioner can readily make diagnoses; but others may find it less easy. In fact, the dream of expert systems work has often been that of the 'expert on the screen': a computer program that can be switched on whenever and wherever expertise is required.
But how would you make such a program?
Surely, it would be difficult?
Medical practitioners may have expertise in their own field: but would they really want to spend years more of their time acquiring significant expertise in the development of computer programs simply in order to give that expertise to others?
And the same is true for other experts in many other fields as diverse as diagnosing disease in humans, or assessing job candidate suitability, or presenting horticultural advice for gardeners.
In general, the answer to this problem is an Expert System Shell.
The reason why programs such as these are called shells is because, of themselves, they contain no specific expertise at all. Instead, they are simply hollow shells into which specific expertise can readily be placed; and in which that expertise can later be accessed and manipulated by a user.
At the outset, the shell is not expert in any specific field whatsoever. It is simply the means by which expert systems can be created.
But, in the past, even expert systems shells have been difficult and demanding to use - they have required considerable expertise in computer science simply in order to understand how to embody the specific expertise that you, personally, might want to encapsulate in your own computerised expert system.
There is now an answer to this problem. The answer is XMaster.
(c) Copyright Chris Naylor 2020