Documentation requirements of the
ANS Forth Standard X3J14

Updates to this file reflect Version 6.08.00, although this is a work in progress.

This file is meant to satisfy the documentation requirements of the ANS Forth Standard X3J14 with regard to the associated Forth system named WIN32FOR, sometimes referred to as Forth-95. It is the intent of the originators of WIN32FOR that it conform to the ANS Forth Standard, including all of the optional wordsets and extensions thereof. All wordsets except the BLOCK and BLOCK-EXT wordsets are available when the Forth system is invoked. The method for obtaining the BLOCK and BLOCK-EXT wordsets is specified at the end of this document, as required by the ANS Forth Standard. If you find that the system and/or the related documentation is incorrect or does not meet the ANS Standard, please let us know.

WIN32FOR is designed to operate in a Windows 2003, XP, 2000, NT4, NT3.51, 98 or 95 environment. Certain features may not operate identically across all operating systems, but these features are outside the ANS Forth domain.

The system has been placed in the public domain, with certain restrictions on the use of the assembler (see 486ASM.DOC). As a public domain system, there is NO support by the authors or subsequent contributors.

Implementation-defined Options

Sections 4.1.1,,,,,,,, and of the ANS Forth Standard require that the system shall document the values for, or behaviours of, each of the following items:

4.1.1 Implementation-defined options

There is no requirement that addresses be aligned. Note, however, that address supplied to certain Windows functions may be required to be aligned on double word boundaries. Note that the data fields of words created by CREATE , arrays and variables start on double word boundaries. Performance may be severely impacted if words, doublewords and quadwords expected by some addressing modes are not aligned to 2, 4 or 8 byte boundaries.

For graphics 0 through 6, 9, 11, 12, 14 through 31, and 127 the default black vertical bar is usually displayed. Graphic character 7 produces an audible short tone or "beep". Graphic character 8 is the "backspace" character. If the cursor is at the left-most position on a line, 8 EMIT will effectively be a no-operation. Otherwise, the previous character is effectively deleted by the sequence backspace, space, and backspace. Graphic character 10 is the line-feed character. 10 EMIT will cause the cursor to be advanced by one line without changing the horizontal position. Graphic character 13 is the "carriage-return" character. 13 EMIT will cause the cursor to be placed at the beginning of the current line. Note that for the purposes of EMIT, 9 (Horizontal Tab) has no special effect. Characters above 127 are generally system dependent.

Editing support is provided for strings up to 126 characters in length, with full word and character operations using keypad or WordStar keys as follows:


Left word


Left character


Right character


Right word


Forward delete


Word delete


Line delete or clear

Left arrow  

Left character

Ctrl-Left arrow  

Left word

Right arrow  

Right character

Ctrl-Right arrow  

Right word


Beginning of line


End of line


Discard changes and leave


Save changes and leave

Characters are usually 8 bits wide in the current implementation. The character set for EMIT depends on the (selectable) current font. The default font is the SYSTEM raster font. See the word SET-FONT in file SCRNCTRL.F. The character set for KEY is system dependent. The use of the cursor and function keys, with or without the SHIFT or CTRL keys may result in values exceeding 255. It is possible to obtain ANSI character values in the range 1 to 255 from KEY by holding the ALT key and typing up to 4 digits on the numeric keypad, beginning with a leading "0" . Thus ALT 0250 will yield the value of 250. Normally one should avoid the use of the ALT key, except for Windows functions.

There are no particular requirements other than that the address must be in a valid range. Each character-aligned address contains 8 bits (1 byte).

Characters above 127 in ASCII equivalents are not modified for matching characteristics, regardless of the state of CAPS .

All characters less than or equal to 32 match a space character when a string is scanned by WORD . Implementation-defined options Ambiguous conditions Other system documentation Implementation-defined options Implementation-defined options Implementation-defined options Implementation-defined options Implementation-defined options Implementation-defined options Implementation-defined options Implementation-defined options

Any restrictions a multiprogramming system places on the use of buffer addresses

The system includes an optional multiprogramming capability. There are no restrictions on the use of any buffer addresses, but inappropriate use may caused unexpected results. In particular, the system does not attempt to serialise their use.

The number of blocks available for source text and data

Depends on the size of the open block file. Method for obtaining the BLOCK and BLOCK Extension wordsets:


Control-flow items are placed on the return stack.

Conversion of digits larger than 35 ( Digit conversion)
Correspondance continues up through character 254, or digit 189.

Display after input terminates in 6.1.0695 ACCEPT and 6.2.1390 EXPECT
No special display to indicate end of input.

Exception abort sequence (as in 6.1.0680 ABORT")
The data stack, return stack and local parameter stack are cleared. QUIT is executed.

Input line terminator ( User Input Device)
The input terminator is the carriage return (character 13).

Maximum size of a counted string, in characters ( Counted strings, 6.1.2450 WORD)
The maximum size of a counted string is 255 characters.
  1. Maximum size of a parsed string, in characters (3.4.1 Parsing)
    The maximum size of a parsed string is 255 characters.
  2. Maximum size of a definition name, in characters ( Definition names)
    The maximum size of a definition name is 255 characters.
  3. Maximum string length for 6.1.1345 ENVIRONMENT?, in characters
    The maximum length for ENVIRONMENT? is 255 characters.
  4. Method of selecting User input device
    Not normally selectable. The words KEY and ACCEPT are deferred words, which may be changed, but it is not a trivial exercise.
  5. Method of selecting User output device
    Not normally selectable. The words EMIT and TYPE are deferred words which may be redirected to serve other output devices.
  6. Methods of dictionary compilation (3.3 The Forth dictionary)

For normal Forth words, a header consists of  (1) a string of characters giving the name of the word, (2) a one byte count field (with high order 2 bits reserved for system use, (3) a 32 bit "view" field (giving information on the source), (4) a 32 bit "link" field pointing to a previous header, (5) a 32 bit execution address field which points to code, and (6) the body of the word which may hold data, addresses, actual code, or a high level definition.

Other words may optionally precede the (1) string of characters naming the word. These fields are documented in the kernel source.

(20) Number of bits in one address unit ( Addresses): The number of bits in one address unit is 8.

(21) Number representation and arithmetic ( Internal number representation: For CONSTANT , VARIABLE , VALUE , members of a LONG-ARRAY (and associated variants), and values stored in memory using ! , the representation is 32 bit 2's complement binary in Little Endian format (least significant 8 bits at the lowest address).

For 2CONSTANT , 2VARIABLE , members of a DOUBLE-ARRAY and values stored in memory using 2! , the representation is 64 bit 2's complement binary with the most significant 32 bits stored at the least significant address in Little Endian format, and the least significant 32 bits stored in Little Endian format starting 4 address units higher then the starting address.

For members of a WORD-ARRAY and values stored in memory using W! , the representation is 16 bit unsigned binary. When the associated value is retrieved using W@, the most significant 16 bits of the item on the data stack are set to zero, without sign extension.

For members of BYTE-ARRAY and values stored in memory using C! , the representation is 8 bits unsigned binary. When the values are retrieved using C@ , the most significant 24 bits of the stack item are set to zero.

(22) Ranges for n, +n, u, d, +d, ud (3.13 Single-cell types, 3.1.4 Cell-pair types):

Type Range

n -2147483648 to 2147483647

+n 0 to 2147483647

u 0 to 4294967295

d -9223372036854775808 to 9223372036854775807

+d 0 to 9223372036854775807

ud 0 to (2**64)-1 = 18446744073709551615

(23) Read-only data-space regions (3.3.3 Data space): No area within Forth is read-only. Areas outside of Forth are read-only as determined by Windows.

(24) Size of buffer at 6.1.2450 WORD ( Other transient regions): The buffer returned by WORD is named POCKET . It is 255 bytes in length. It does not overlap the buffer returned by #> .

(25) Size of one cell in address units (3.1.3 Single-cell types): One cell encompasses 4 address units.

(26) Size of one character in address units (3.1.2 Character types): One character encompasses 1 address unit.

(27) Size of the keyboard terminal input buffer ( Input buffers): The terminal input buffer is 255 bytes long

(28) Size of the pictured numeric output string buffer ( Other transient regions): The pictured numeric output string buffer is 80 bytes long.

(29) Size of the scratch area whose address is returned by 6.2.2000 PAD ( Other transient regions): The size of the scratch area returned by PAD is at least 255 bytes.

(30) System case-sensitivity characteristics (3.4.2 Finding definition names): The system default is case-insensitive, as determined by the truth value in the variable CAPS. If the variable is set to true (the default value), lower case characters (letters a through z) are converted to upper case before any dictionary searches. Furthermore, the names of words added to the dictionary will be in upper case. If CAPS is set to false, characters will not be changed prior to dictionary searches or before being placed in the dictionary.

(31) System prompt (3.4 The Forth text interpreter, 6.1.2050 QUIT): The system prompt consists of a space, the letters ok and a line feed and carriage return.

(32) Type of division rounding ( Integer division, 6.1.0100 */, 6.1.0110 */MOD , 6.1.0230 / , 6.1.0240 /MOD , 6.1.1890 MOD): Floored division is used in the operators */ */MOD / /MOD and MOD .

(33) Values of 6.1.2250 STATE when true: When STATE is true, its value is -1 (all bits set).

(34) Values returned after arithmetic overflow ( Other integer operations): For addition, subtraction, and multiplication arithmetic overflow will generally be the result of an infinitely long 2's complement result truncated to the number of bits allowed in the result of the operation. The result of arithmetic overflow in division or modulus operations is not readily predictable.

(35) Whether the current definition can be found after 6.1.1250 DOES> (6.1.0450 : ) The name of the current definition can not be found after the occurrence of the word DOES> until the terminating ; is encountered.

(36) The format used for display by LIST (if implemented): Assuming that the default values in BLOCK.F are use, LIST will cause the printing of 16 lines of 64 characters each from the current block file. There is no explicit numbering of lines and no explicit display of the end of a line.

(37) The length of a line affected by \ (if implemented): Using the default values of C/L , the length of a line affected by \ while using \ in a block file is 64 characters, including the character \ itself, assuming that it is the first character on the line.

(38) Values used in the system by CATCH and THROW (9.3.1 THROW values, 9.3.5 Possible actions on an ambiguous condition): The following values used by THROW are the only ones currently implemented:

-1 (for ABORT ) -2 (for ABORT" ) -3 (for stack overflow) -4 (for stack underflow) -13 (for undefined words)

(39) Encoding of keyboard events ( EKEY): This may be dependent on the particular system and keyboard used. For the system and keyboard used for initial testing of this system, the standard keys and control key combinations will be returned. In addition the following hexadecimal values will be returned:

Hex value Key

0x0000003 Ctl Break 0x0010001 F1 0x0010002 F2 . . . . 0x0010009 F9 0x0020000 Home 0x0020001 End 0x0020002 Ins 0x0020003 Del 0x0020004 Left arrow 0x0020005 Right arrow 0x0020006 Up arrow 0x0020007 Down arrow 0x0020008 Scroll Lock 0x0020009 Pause 0x0020010 Pg Up 0x0020011 Pg Dn 0x0050001 Ctl F1 . . . . 0x0050009 Ctl F9 0x0090001 Shft F1 . . . . 0x0090009 Shft F9 0x00A0009 Shft Break 0x00D0001 Ctl shft F1 . . . . 0x00D0009 Ctl shft F9 0x0200001 Left mouse button 0x0200002 Right mouse button

It is also possible to enter numeric ASCII values directly with the ALT key in conjunction with the numeric keypad. The ASCII value should be in the range of 1 to 255, and must be preceeded with a 0.

(40) Duration of a system clock tick: 1 millisecond.

(41) Repeatability to be expected from execution of MS: 20 milliseconds.

(42) File access methods used by BIN, CREATE-FILE, OPEN-FILE, R/O, R/W, and W/O: The values of file access methods allowed are the following:

r/o 0x80000000 r/w 0xc0000000 w/o 0x40000000

(43) Format and range of floating-point numbers (12.3.1 Data types, REPRESENT): Default floating-point numbers are represented in the 64 bit IEEE 754 format, in which the most significant bit is the sign bit, the next 11 bits are the exponent bits (biased by 1023), and the remaining 52 bits represent the fractional binary bits of the significand. There is an implied bit of 1 to the left of the 52 bits, which normally represents the integer part of the significand. The approximate numerical range of represented numbers is 2.23 x 10**-308 to 1.79 x 10**308. Other values represented include +0, -0, +infinity, -infinity, and various forms of "Not a Number" or "NaNs". The 32 bit IEEE 754 format is also avaliable using the operators SF@ and SF!. An 80 bit format may be made available, but requires changing the internal constant B/FLOAT in the file FLOAT.F to have the value 10 , and then re-compiling the system. Note that the floating-point numbers appear in memory in Little-Endian format, with the least significant 8 bits at the lowest address.

(44) Results of REPRESENT when float is out of range: The results of REPRESENT when float is out of range depend on the particular value being represented. The character count passed to REPRESENT is the maximum of one of the following character strings:

"Unsupported" "NAN" "Infinity" "Empty" "Unknown"

The "Unsupported" string is returned when the floating-point number has a representation not supported by the Intel Numeric Processing Unit (NPU). "NAN" is returned when the floating-point number has one of the "Not-a-Number" values specified by the IEEE 754 Standard. Similarly, "Infinity" is returned when the floating-point value has one of the values of infinity specified by the IEEE 754 Standard. The "Empty" string is returned when the floating-point stack is empty. The "Unknown" string is returned when the floating- point value has an unknown format (normally seen only in the 10 byte mode).

(45) Rounding or truncation of floating-point numbers ( Floating-point numbers): Floating point numbers are rounded to the nearest represent- able value. If the infinitely precise value is 1/2 the least significant representable bit, the result is rounded so that the least significant bit is 0 (the "round to even" rule in the IEEE 754 Standard.

(46) Size of floating-point stack (12.3.3 Floating-point stack): The floating-point stack holds up to 256 floating-point numbers.

(47) Width of floating-point stack (12.3.3 Floating-point stack): The default floating-point stack is 8 bytes wide. As mentioned earlier, this may be changed by recompilation.

(48) Maximum number of locals in a definition (13.3.3 Processing locals, LOCALS|): The maximum number of local variables is 12.

(49) Values and meaning of ior (14.3.1 I/O Results data type, ALLOCATE, FREE, RESIZE): For ALLOCATE and RESIZE , ior returns the value TRUE (-1) on an error and FALSE (0) for no errors. FREE always returns a FALSE value.

(50) Ending sequence for input following ;CODE and CODE : END-CODE

(51) Manner of processing input following ;CODE and CODE : Assuming that the assembler has been loaded (the default condition), lines of assembly code source are assumed to follow. The ASSEMBLER vocabulary is added to the vocabulary search order and the code is assembled until the word END-CODE is encountered. The ASSEMBLER vocabulary is then removed from the vocabulary search order.

(52) Search-order capability for EDITOR and ASSEMBLER (15.3.3 The Forth dictionary):

(53) Maximum number of word lists in the search order (16.3.3 Finding definition names, SET-ORDER): The maximum number of word lists in the search order is 12.

General ambiguous conditions (4.1.2)

A system shall document the system action taken upon each of the general or specific ambiguous conditions identified in this Standard. See 3.4.4 Possible actions on an ambiguous condition. The following general ambiguous conditions could occur because of a combination of factors:

(1) A name is neither a valid definition name nor a valid number during text interpretation (3.4 The Forth text interpreter); If the error is not an object or a class, display an error message:

Error: <name> is undefined

Otherwise, if the name consists of <chars>.<chars> or <chars>:<chars> (where <chars> is one or more characters) find returns a word that finds the name at run-time.

(2) A definition name exceeded the maximum length allowed ( Definition names): Display an error message:

| Truncating name longer than 255 characters! | -> <name> | | The first 255 characters of the name are retained as the | name of the definition.

(3) Addressing a region not listed in 3.3.3 Data Space: The message EXCEPTION 0xC0000005 ACCESS_VIOLATION is displayed, followed by a dump of the registers and the return stack.

(4) Argument type incompatible with specified input parameter, e.g., passing a flag to a word expecting an n (3.1 Data types): Ignore and continue.

(5) Attempting to obtain the execution token, (e.g., with 6.1.0070 ', 6.1.1550 FIND, etc.) of a definition with undefined interpretation semantics: Display an error message:

Error: <name> is undefined

(6) Dividing by zero (6.1.0100 */, 6.1.0110 */MOD, 6.1.0230 /, 6.1.0240 /MOD, 6.1.1561 FM/MOD, 6.1.1890 MOD, 6.1.2214 SM/REM, 6.1.2370 UM/MOD, M*/): The message EXCEPTION 0xC0000094 INT_DIVIDE_BY_ZERO is displayed, followed by a dump of the registers and the return stack.

(7) Insufficient data-stack space or return-stack space (stack overflow): The application terminates silently.

(8) Insufficient space for loop-control parameters: The application terminates silently.

(9) Insufficient space in the dictionary: Display an error message:

Error: ALLOT Out of memory!

(10) Interpretating a word with undefined interpretation semantics: Display an error message:

Error: <name> compilation only

(11) Modifying the contents of the input buffer or a string literal ( Text-literal regions, Input buffers): No explicit system action (the action itself will not be detected, but may result in unwanted and unstandard actions.

(12) Overflow of a pictured numeric output string: May cause unspecified error.

(13) Parsed string overflow: String is clipped to a length of 255.

(14) Producing a result out of range, e.g., multiplication (using *) results in a value too big to be represented by a single-cell integer (6.1.0090 *, 6.1.0100 */, 6.1.0110 */MOD, 6.1.0570 >NUMBER, 6.1.1561 FM/MOD, 6.1.2214 SM/REM, 6.1.2370 UM/MOD, 6.2.0970 CONVERT, M*/): Ignore and continue.

(15) Reading from an empty data stack or return stack (stack underflow): Display an error message:

Error: stack underflow

(16) Unexpected end of input buffer, resulting in an attempt to use a zero-length string as a name: In most cases, display an error message.

Specific ambiguous conditions:

The following specific ambiguous conditions are noted in the glossary entries of the relevant words:

(1) >IN greater than size of input buffer (3.4.1 Parsing): WORD will return an undefined string.

(2) 6.1.2120 RECURSE appears after 6.1.1250 DOES> : Ignore and continue.

(3) Argument input source different than current input source for 6.2.2148 RESTORE-INPUT : Will cause an undefined system behavior.

(4) Data space containing definitions is de-allocated ( Contiguous regions): Ignore and continue.

(5) Data space read/write with incorrect alignment ( Address alignment): Ignore and continue.

(6) Data-space pointer not properly aligned (6.1.0150 ,, 6.1.0860 C,): Ignore and continue.

(7) Less than u+2 stack items (6.2.2030 PICK, 6.2.2150 ROLL): Ignore and continue.

(8) Loop-control parameters not available (6.1.0140 +LOOP, .1.1680 I, 6.1.1730 J, 6.1.1760 LEAVE, 6.1.1800 LOOP, 6.1.2380 UNLOOP): Ignore and continue.

(9) Most recent definition does not have a name (6.1.1710 IMMEDIATE): Invert the state of the "immediate" bit in the header of the most recently defined word with a header.

(10) Name not defined by 6.2.2405 VALUE used by 6.2.2295 TO : The message EXCEPTION 0xC0000005 ACCESS_VIOLATION is displayed, followed by a dump of the registers and the return stack.

(11) Name not found (6.1.0070 ', 6.1.2033 POSTPONE, 6.1.2510 ['], 6.2.2530 [COMPILE] ): Display an error message:

Error: <name> is undefined

(12) Parameters are not of the same type (6.1.1240 DO, 6.2.0620 ?DO, 6.2.2440 WITHIN): Ignore and continue.

(13) 6.1.2033 POSTPONE or 6.2.2530 [COMPILE] applied to 6.2.2295 TO : Ignore and continue.

(14) String longer than a counted string returned by 6.1.2450 WORD: WORD will return a length not exceeding 255, and skip over any remaining characters.

(15) u greater than or equal to the number of bits in a cell ( 6.1.1805 LSHIFT, 6.1.2162 RSHIFT): Only the least significant 5 bits of the count u are used in the shift count.

(16) Word not defined via 6.1.1000 CREATE (6.1.0550 >BODY, 6.1.1250 DOES> ): Ignore and continue.

(17) Words improperly used outside 6.1.0490 <# and 6.1.0040 #> (6.1.0030 #, 6.1.0050 #S, 6.1.1670 HOLD, 6.1.2210 SIGN ): Ignore and continue.

(18) Correct block read was not possible: Display an error message:

Error: Error reading block

(19) I/O exception in block transfer: Display one of the following error messages:

Error: Error reading block Error: Error writing block, probably out of disk space.

(20) Invalid block number ( BLOCK, BUFFER, LOAD ): Display an error message:

Error: Error reading block

(21) A program directly alters the contents of BLK : Ignore and continue.

(22) No current block buffer for UPDATE : Block 0 of the current block file will be updated.

(23) d outside range of n in D>S : The high order part of the number is discarded. Any errors are ignored.

(24) AT-XY operation can't be performed on user output device: Ignore and continue.

(25) Attempting to position a file outside its boundaries ( REPOSITION-FILE): Ignore and continue.

(26) Attempting to read from file positions not yet written ( READ-FILE, READ-LINE ): A count of 0 is returned.

(27) fileid is invalid ( INCLUDE-FILE ): A message box labeled "Exception" is displayed with the exception number "C0000005". When the "OK" is chosen, the message "EXCEPTION: ACCESS_VIOLATION" is displayed.

(28) I/O exception reading or closing fileid ( INCLUDE-FILE, INCLUDED ): Errors encountered on closing a file are ignored. If an error occurs during reading, display an error message:

Error: read error

(29) Named file cannot be opened ( INCLUDED ): Display an error message:

Error: <filename> file not found

(30) Requesting an unmapped block number (11.3.2 Blocks in files): Display an error message:

Error: Error reading block

(31) Using SOURCE-ID when BLK is not zero: Not an error.

(32) DF@ or DF! is used with an address that is not double-float aligned: This is not an error: all addresses are double-float aligned.

(33) F@ or F! is used with an address that is not float aligned: This is not an error: all addresses are double-float aligned.

(34) Floating point result out of range (e.g., in F/ ): If the result would overflow, a result of "infinity" is returned.

(35) SF@ or SF! is used with an address that is not single-float aligned: This is not an error: all addresses are single-float aligned.

(36) BASE is not decimal ( REPRESENT, F., FE., FS. ): REPRESENT returns the proper values with the string having decimal digits. F. displays the value in decimal. FE. and FS. display in decimal.

(37) Both arguments equal zero ( FATAN2 ): A floating point 0 is returned.

(38) Cosine of argument is zero for FTAN : There is no representable floating point value for which the cosine is exactly zero.

(39) d can't be precisely represented as float in D>F : In the default 64 bit floating-point mode, the double precision value is rounded to 54 bits of significance, using the least most 10 bits rounded to the nearest even value.

(40) Dividing by zero ( F/ ): The result is "Infinity", and there is no error reported.

(41) Exponent too big for conversion ( DF!, DF@, SF!, SF@ ): In the default 64-bit precision mode, the representations for floating point numbers is the same as used in D@ and D! . For those cases there is no conversion error. For SF@ the conversion will be exact. For SF! with a positive exponent too large for conversion, the resulting value is "infinity". For SF! with a negative exponent too large in magnitude, the result is zero.

(42) float less than one ( FACOSH ): No error is reported. The function is calculated using absolute values where appropriate to avoid errors.

(43) float less than or equal to minus-one ( FLNP1 ): For a value of exactly minus-one, the result is "-infinity". For values less than minus-one, the absolute value of the value plus one is taken prior to the logarithm function. No error is reported.

(44) float less than or equal to zero ( FLN, FLOG ): No error is reported. For a value of zero, the result is "-infinity". For negative values, the absolute value is applied prior to taking the logarithm function.

(45) float less than zero ( FASINH, FSQRT ): No error is reported. The absolute value is applied prior to performing the function.

(46) float magnitude greater than one ( FACOS, FASIN, FATANH ): No error is reported.

(47) integer part of float can't be represented by d in F>D : The result is the most negative 64 bit double number. No error is reported.

(48) String larger than pictured-numeric output area ( F., FE., FS. ): Not an error condition.

(49) Executing a named local while in interpretation state ( (LOCAL) ): The named local will be undefined.

(50) Name not defined by VALUE or LOCAL ( TO ): The result is unspecified.

(51) Deleting the compilation word-list ( FORGET ): Not possible.

(52) Fewer than u+1 items on control-flow stack ( CSPICK, CSROLL ): The result is undefined.

(53) Name can't be found ( FORGET ): Display an error message:

Error: <name> is undefined

(54) Name not defined via 6.1.1000 CREATE ( ;CODE ): Display an error message:

Error: ALLOT Out of memory!

(55) 6.1.2033 POSTPONE applied to [IF] : [IF] compiled into definition.

(56) Reaching the end of the input source before matching [ELSE] or [THEN] ( [IF] ): Display an error message:

Error: Can't refill the input stream.

(57) Removing a needed definition ( FORGET ): Not an error: unhooks any forgotten deferred word back to its default behaviour.

(58) Changing the compilation word list (16.3.3 Finding definition names): No effect on the compiler.

(59) Search order empty ( PREVIOUS ): PREVIOUS doesn't allow the wordlist to be empty: if it would be empty, it inserts the ROOT vocabulary.

(60) Too many word lists in search order ( ALSO ): The least recently added vocabulary or ROOT is discarded.

Other system documentation required by the ANS Standard (Section 4.1):

A system shall provide the following information:

(1) A list of non-standard words using 6.2.2000 PAD ( Other transient regions): None.

(2) Operator's terminal facilities available: Full line editing including keypad cursor controls and Wordstar functions ^A, ^S, ^D, ^F, ^G, ^T, ^E, and ^X.

(3) Program data space available, in address units: 199980 address units, in the default system, as distributed. If more space is desired, another Forth system may be created by specifying a larger number with the MEMORY-FREE! operator, then saving the Forth system using FSAVE . For an example, see the end of the file EXTEND.F .

(4) Return stack space available, in cells: At least 4095 cells.

(5) Stack space available, in cells: At least 4095 cells.

(6) System dictionary space required, in address units: 512,000 address units.

Document $Id: p-ansi.htm,v 1.2 2005/06/27 15:42:21 dbu_de Exp $