D O C T O R  (THE COMPUTERIZED ANALYST)

ABSTRACT:

	THE PROGRAM "DOCTOR" IS A VERSION OF THE OLD "ELIZA" PROGRAM
WRITTEN TO RUN IN TECO. IT IS LESS POWERFUL THAN THE ORIGINAL LISP
PROGRAM, BUT REQUIRES ONLY A VERSION OF TECO TO RUN. IT CAUSES THE
COMPUTER TO RESPOND TO YOUR STATEMENTS AS WOULD (?) A PSYCHO-
ANALYST, AND ATTEMPTS TO CARRY OUT A CONVERSATION WITH THE USER.


PROGRAM DESCRIPTION:

	DOCTOR IS IMPLEMENTED IN 4 FILES:
DOCTOR.TEC	IS THE COMMAND FILE (I.E., THE "PROGRAM")
DOCTOR.KEY	IS THE LIST OF KEY WORDS/PHRASES LOOKED FOR IN YOUR COMMENTS
DOCTOR.NUM	IS THE LIST OF OFFSETS INTO ITS REPLY LIST OF REPLIES
DOCTOR.RPL	IS THE LIST OF CANNED REPLIES TO YOUR QUESTIONS.

	DURING EXECUTION, DOCTOR.KEY RESIDES IN Q-REGISTER K, DOCTOR.NUM
RESIDES IN Q-REGISTER N, DOCTOR.RPL RESIDES IN Q-REGISTER R, AND BY
CONVENTION (THOUGH IT NEED NOT BE SO), DOCTOR.TEC RESIDES IN Q-REGISTER Z.
IF YOUR TECO DOES NOT HAVE ENOUGH BUFFER SPACE TO HOLD ALL THESE (PLUS
A SECOND COPY OF THE REPLY LIST, WHICH IS THE LARGEST), YOU NEED TO ADD
SPACE SOMEHOW TO USE THIS PROGRAM, OR RE-WRITE IT ENOUGH SO THAT
IT SEEKS REPLIES OFF DISK INSTEAD OF BY LOADING Q-REGISTER R FOR
REPLIES. (THE OTHER LISTS ARE SHORT ENOUGH THAT THEY SHOULD CAUSE
LITTLE TROUBLE.) IN RSX11D, A TECO WITH 8000 WORDS OF STACK IS BIG
ENOUGH; WITH 6000 IT MAY BE, BUT IS NOT TRIED.
WITH OTHER SYSTEMS, GOOD LUCK! DOCTOR SHOULD RUN PRETTY MUCH INDEPENDENTLY
OF THE OPERATING SYSTEM OR TECO DIALECT.

	THE PROGRAM IS REALLY VERY SIMPLE IN ITS OPERATION. IT WORKS ON
YOUR INPUT LIKE THIS:

1. READ IN A SENTENCE OF TEXT FROM THE USER, REMOVING ANY "'" CHARACTERS
	AS THEY ARE READ IN. UPON RECEIPT OF CARRIAGE RETURN, STORE THE
	LINE AND PROCEED.
2. PULL IN ONE LINE OF THE KEYWORD LIST AT A TIME AND SEE IF ANY OF ITS
	STRINGS ARE FOUND IN THE SENTENCE READ. CONTINUE UNTIL EITHER
	A KEYWORD IS FOUND OR THE LIST IS EXHAUSTED.
3. STORE THE NUMBER OF LINES INTO THE KEYWORD LIST YOU HAVE GONE, AND
	SAVE FOR LATER.
4. STORE THE PART OF THE SENTENCE READ THAT IS AFTER THE KEYWORD FOUND.
5. TRANSLATE PERSONAL PRONOUNS (E.G. "YOU" TO "I", "YOUR" TO "MY", ETC.)
	IN THE SENTENCE PART STORED IN STEP 4 SO THAT THEY CAN BE USED
	IN POSSIBLE REPLIES.
6. LOAD THE LIST OF NUMERIC OFFSETS AND SKIP DOWN THE SAME NUMBER OF LINES
	SKIPPED TO FIND THE KEYWORD. CONVERT THE NUMBERS TO BINARY AND
	SAVE THEM. THE FIRST IS THE NUMBER OF LINES INTO THE REPLY
	BLOCK THE FIRST REPLY FOR THIS KEYWORD IS, AND THE SECOND IS THE
	NUMBER OF REPLIES FOR THIS KEYWORD.
7. LOAD THE LIST OF REPLY STATEMENTS. SKIP DOWN BY THE NUMBER OF LINES TO
	THE FIRST REPLY FOR YOUR KEYWORD. THEN SKIP DOWN BY THE NUMBER
	OF SENTENCES READ, MODULO THE NUMBER OF REPLIES ALLOWED FOR THIS
	KEYWORD. (THUS, YOU SKIP AN EXTRA 0 TO N-1 LINES, WHERE THERE ARE
	N LINES ALLOWED FOR THIS KEYWORD.)
8. SAVE THE REPLY STATEMENT AND ERASE THE REST OF THE BLOCK.
9. IF THE REPLY CONTAINS A "*" CHARACTER, REPLACE IT WITH A SPACE FOLLOWED
	BY THE TRANSLATED PART OF THE ORIGINAL SENTENCE AFTER THE KEYWORD.
10. TYPE IT OUT AND GO ON.

	THIS CONTINUES UNTIL "GOODBYE" IS SEEN, AT WHICH POINT A CLOSING
MESSAGE IS TYPED AND TECO EXITS.