			Addenda, Emenda & Corrigienda
				   to
			  _Forth: The New Model_
			   (c) 1994 Jack Woehr
			       PO BOX 51
			   GOLDEN CO 80402-0051
			     jax@well.sf.ca.us
			 SYSOP, RCFB (303) 278-0364

( Revision $Revision: 1.5 $)
( Corrections welcome, email me please.)

!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
THE AUTHOR DOES NOT WARRANT THIS DOCUMENT IN ANY FASHION.
!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

This document provides a number of additions, amendments and corrections
to the book _Forth: The New Model_ by Jack Woehr, M&T Books, 1992, ISBN
1-55851-277-2. This information is provided as a courtesy to the readers
and purchasers of that book. This document may be freely redistributed
where such redistribution does not conflict with the law or the rights
of others.

Errors noted herein have been selected according the the following pair
or criteria, and only the following pair:

	A) I have noticed or a reader has noticed the error

		&&AND&&

	B) I find the error significant enough to warrant inclusion
	   in this document.

Ditto emenda, ditto addenda.

UPDATES
-------
As people point out more errors to me, I'll try to update and repost
this document, which will be found on the RealTime Control & Forth
Board (RCFB) (303) 278-0364 and selected ftp sites in the file
4THNUMOD.ZIP.

AMERICAN NATIONAL STANDARD FORTH
--------------------------------
American National Standard Forth is copyrighted by the American National
Standards Institute, 11 West 42nd Street, New York, NY 10036. For
information on how to obtain the Standard document, please contact that
organization, or alternatively, contact the X3J14 Technical Committee by
emailing ansforth-request@minerva.com via Internet email.

/////////////////////////////
ADDENDA, EMENDA & CORRIGIENDA
/////////////////////////////

	Some errors are corrected here. Some changes to the draft
Standard are noted which occurred subsequent to the writing of the book.
Some additions to the Standard are noted.

	The changes in the Standard have not served to invalidate any
code examples in the book. In a single case (see below), a code example
in the book was originally incorrect as printed (page 107 of the book).

Page	Information
----	-----------
 11	The last draft proposed Standard, d-pANS-6, was approved
	recently by the X3 body overseeing X3J14's work. Elizabeth
	Rather mentioned this in electronic mail on April 19th:

	"I'm happy to inform you that, according to Lynn Barra
	of X3, the ANSI BSR letter ballot for final approval of ANS
	Forth passed.  Lynn believes they will take about 2 mos. to
	publish the document, and will issue a press release when it
	is ready."

 65	The stack diagram and example for FIND are incorrect. The correct
	stack diagram is on page 267:

		FIND ( c-addr -- c-addr 0 | xt 1 | xt -1)

	The FIND example with DUP should start off

		C" DUP" FIND

	since FIND expects the count character address of a counted
	string. If not found find, that same address of the count
	character of a counted string will be returned with zero (0)
	on top of the data stack; otherwise, the execution token is
	returned with the 1 flag for IMMEDIATE words and -1 for non-
	IMMEDIATE words. (C" is on page 139.)

	Note that

		S" DUP" DROP -1 CHARS + FIND

	would not be correct (though it might work on some systems)
	since S" is only guaranteed to return a count and an address,
	the format of the storage used by S" being undefined.

 82	The explanation of the results of FM/MOD and SM/REM is
	mathematically correct but in one instance in conflict with
	the Standard and with Forth tradition! The result of

		-7. -3 FM/MOD

	is actually the same as SM/REM, that is

		2 -1

	which happens to be what Jax4th returns correctly, so don't
	ask me what I was thinking when I wrote the erroneous example!

107	The assertion in the GIANT-STEPS example that a DO .. +LOOP
	will go "right past our stop" is erroneous. The first time
	that the unsigned comparison is made and the outer loop index
	is found to have been passed, the loop will exit properly.

109	WITHIN is explained poorly. WITHIN manipulates a circular number
	space. In a 16-bit system, -7FFE is WITHIN 7FFF and -7FFD. The
	Standard document q.v. explains this phenomenon more precisely.

123	D@ and D! vanished from the Standard.

137	I refer to "Microsoft Unicode". Although Microsoft was a major
	driving force and early implementor of Unicode, Unicode is the
	product of the Unicode Consortium, which includes many other
	vendors and parties interested in internationalization. See
	_The Unicode Standard: Worldwide Character Encoding, Version 1.0_,
	Volumes 1 and 2, Addison-Wesley Publishing Company, 1991, 1992,
	ISBN 0-201-56788-1. 

159	TIB and #TIB have been "banished" to CORE EXTENSIONS. Added to
	the CORE is the word SOURCE ( -- c-addr u) which provides unified
	access to the input stream under all 	circumstances. At last!
	Some of us fought for this for years. See the Standard for more
	details.

170	SAVE-INPUT and RESTORE-INPUT have changed slightly in detail,
	conceptually occupy the same space as noted in the book.

182	Table 10-1 has changed somewhat in the final draft of the
	Standard.

183	The specification of CATCH and THROW has been tightened up
	somewhat, but is conceptually the same as in the book. All
	code examples in the book will still function.

224	The discussion of centennial leap years under Chapter 5,
	Example 3 is backwards: a centennial leap year occurs in
	every centennial year evenly divisible by 400, and not in
	the other three centennial years that epoch. 2000 is indeed
	a leap year; 1900 was not.

236	Dick and Jill Miller's company is "Miller Microcomputer
	Services", not "Miller Micro Systems", 61 Lake Shore Road,
	Natick, MA 01760-2099. Their phone is (508) 653-6136, call
	(9am-9pm Eastern Time Zone). Their Internet address is
	dmiller@im.lcs.mit.edu. Incidentally, they have a lot
	of hard-to-get and out-of-print Forth literature for sale.

241	Throughout the Glossary, the "C" attribute (indicating
	compile-time only) is used. This attribute marker as a
	notation in the Standard document was removed from the
	later drafts. I view this removal as a minor but tangible
	flaw in the final Standard.

302	_The Computer Journal_'s acronym is TCJ, not TJC.

/////////////
Miscellaneous
/////////////

	The interesting SLITERAL ( c-addr1 u --) was added to the STRING
	wordset. See the Standard for more information.

<<END OF DOCUMENT>>

