2010-08-31  Removed FORGET from the list of words removed, leaving
             it marked as obsolete.
2009-09-09  Removed legacy word set in favour of section D.7
1998-11-30  Original text by Len Zettel.

Problem
=======
Section 1.4.2 "Obsolescent features" declared seven words as
obsolescent and warned "they may be withdrawn from future revisions
of the Standard".  It is now 15 years since that warning was issued,
it is time to remove these words from the document.

In order to do this a number of alterations to the main document
are required.

Proposal
========

1. Replace the third paragraph of section 1.4.2 "Obsolescent
    features", listing the obsolescent words:

       6.2.0060  #TIB      15.6.2.1580  FORGET    6.2.2240  SPAN
       6.2.0970  CONVERT      6.2.2040  QUERY     6.2.2290  TIB
       6.2.1390  EXPECT

    with:

       15.6.2.1580  FORGET

2. Remove reference to EXPECT and CONVERT from the Documentation
    requirements:

    4.1.1 Implementation-defined options:

       - character editing of ACCEPT and EXPECT;

       - display after input terminates in ACCEPT and EXPECT;

    4.1.2 Ambiguous conditions

       - producing a result out of range, e.g., multiplication
         (using *) results in a value too big to be represented by a
         single-cell integer (*, */, */MOD, >NUMBER, FM/MOD, SM/REM,
         UM/MOD, CONVERT, M*/);

    4.2.1 Environmental dependencies

       - depending on the presence or absence of non-graphic
         characters in a received string (ACCEPT, EXPECT);

3. Replace throw code -15:

       -15  invalid FORGET

     with:

       -15  reserved

4. CORE EXT Wordset

    6.1.2216 SOURCE
    Remove ", declaring TIB and #TIB obsolescent" from the end of the
    first paragraph, and the following text from the rationale:

       SOURCE in this form exists in F83, polyFORTH, LMI's Forths
       and others. In conventional systems it is equivalent to the
       phrase

          BLK @ IF BLK @ BLOCK 1024 ELSE TIB #TIB @ THEN

    6.1.2450 WORD
    Remove the sentence "A space, not included in the length,
    follows the string" from the second paragraph and the note
    in the definition.

    Remove the second paragraph from the rationale:

       As a concession to the obsolesecent word CONVERT, Forth94
       required an implementation to put a space after the string
       placed by WORD.  This standard neither requires nor
       prohibits such a space.

    6.2.0060  #TIB
    Remove the definition and rationale.

    6.2.0695 ACCEPT
    Remove reference to EXPECT from the first paragraph, and the
    following paragraphs from the rationale:

       ACCEPT and EXPECT perform similar functions.  ACCEPT is
       recommended for new programs, and future use of EXPECT
       is discouraged.

       As in previous standards, EXPECT returns the input string
       immediately after the requested number of characters are
       entered, as well as when a line terminator is received.
       The "automatic termination after specified count of
       characters have been entered" behavior is widely considered
       undesirable because the user "loses control" of the input
       editing process at a potentially unknown time (the user does
       not necessarily know how many characters were requested from
       EXPECT).  Thus EXPECT and SPAN have been made obsolescent and
       exist in the Standard only as a concession to existing
       implementations. If EXPECT exists in a Standard System it
       must have the "automatic termination" behavior.

       ACCEPT does not have the "automatic termination" behavior of
       EXPECT.  However, because external system hardware and
       software may perform the ACCEPT function, when a line
       terminator is received the action of the cursor, and therefore
       the display, is implementation-defined. It is recommended
       that the cursor remain immediately following the entered text
       after a line terminator is received.

    6.2.0970  CONVERT
    Remove the definition and rationale.

    6.2.1390  EXPECT
    Remove the definition and rationale.

    6.2.2040  QUERY
    Remove the definition and rationale.

    6.2.2125 REFILL
    Remove the reference to QUERY in the rationale leaving:

       REFILL is designed to behave reasonably for all possible
       input sources. If the input source is coming from the user
       REFILL could still return a false value if, for instance, a
       communication channel closes so that the system knows that
       no more input will be available.

    6.2.2240  SPAN
    Remove the definition, there is no rationale.

    6.2.2290  TIB
    Remove the definition and rationale.

5. Remove the reference to #TIB and TIB from the first sentence of
    the "Data space" rationale (A.3.3.3).

6. Remove the reference to CONVERT in the rationale for the
    Core extension words (A.6.2):

    - Words that are being deprecated in favour of new words
      introduced to solve specific problems (e.g., CONVERT).

7. Add the following to Annex D:

    D.7 ANS Forth '94
        =============

    D.7.1 Removed Definitions
          -------------------

    This standard removes six words that were marked 'obsolescent'
    in the ANS Forth '94 document.  These are:

       6.2.0060  #TIB      6.2.1390  EXPECT   6.2.2240  SPAN
       6.2.0970  CONVERT   6.2.2040  QUERY    6.2.2290  TIB

    Reuse of these names is strongly discouraged.

Remarks
=======

Annex D needs work to bring it up to date, let along explain the
difference between Forth200x and ANS Forth '94.  This should be
the topic of another RfD.  The text given in point 9 is intended
as a place holder until annex D is discussed in full.

This proposal leaves FORGET as an obsolescent word.  Another proposal
may be raised to re-instate the word, removing it from the obsolescent
list.

*) If someone really wants to make the effort, they can make a poll on
how many systems implement and how many programmers use FORGET, but I
think that that won't make a difference; it has become pretty clear
that there are a significant number of people who still use FORGET, so
that systems that have it won't remove it.