native.txt

prolog开发工具

ABOUT
-----

Prolog, as is, is very close to a rule language for an expert system.  It
needs a few extra things to make it an expert system.  These include
the ability to explain and the ability to remember answers to questions.

NATIVE.PRO is a shell that provides those added functions.


USE
---

From DOS type:

	c:\> native

From the listener consult either the source file, NATIVE.PRO, or the compiled
module NATIVE.PLM.  To start it, type:

	?- [native].	% equivalent to consult('native.pro').
	?- main.

At the > prompt enter various commands, always ending in . because they
are Prolog terms being read.

help.	Lists the legal commands.

load.	Prompts for the name of the knowledge base to load.  must be a
	Prolog term meaning lots of punctuation. For example 'birds.nkb'.

solve.	Starts a consultation with the loaded knowledge base.

	Answer the questions with either a yes or no or by giving a number
	in answer to menu type questions.

	You can also answer why. to a question.  in this case the list of
	goals trying to be proved is displayed.  After the reason is 
	displayed, you should enter the answer to the question at the > 
	prompt.

how(g). Once the answer is known, how(goal) lets you find out the
	reasoning that led to that goal.  The top goal is birds for birds
	so the command is how(bird(black_footed_albatross)).

	Once the goals are shown, how can be asked again, such as 
	how(order(tubenose)).

whynot(g). Lets you find out why a goal was discarded.

quit.	Exits.


PROJECTS
--------

Fix reads so they don't require periods at the end.

Fix read in load so you can just enter first part of file name - e.g. birds
	-- and you add the .nkb extension automatically.

Fix explanations that are ask or menuask do give a wiser answer like you
	told me or something.

After a why it should reask the question.

Have it lead you through the logical sequence of things - what knowledge
	base do you want to load, then automatically start the consultation.

Have the knowledge base include an intro that is called first - to describe
	the particular knowledge base.

Improve error handling.