reduce-historical/r30/sldec.doc

University of Utah Symbolic Computation Group               July 1978
Technical Report No.  TR-2                First Revision - March 1979










                               MANUAL
                                 for
                            STANDARD LISP
                                 on
                         DECSYSTEM 10 and 20

                                 by

                           Inge B. Frick

                         University of Utah
                      Salt Lake City, UT 84112










                              ABSTRACT



        An implementation of Standard Lisp on  the  DECsystem
        10  and 20 is presented.  The implementation is based
        on Stanford Lisp 1.6.









Work supported in part by the National Science Foundation under Grant
No.  MCS 76-15035.
STANDARD LISP ON DECSYSTEM 10 AND 20


                          Table of Contents
                          _____ __ ________



1. PRELIMINARIES .........................................   1
1.1 Document Conventions .................................   1
1.2 Special Terminal Control Characters ..................   2
1.3 Data Types ...........................................   2
1.4 Property Lists .......................................   5
1.5 Value Cells ..........................................   6
1.6 Function Cells .......................................   6
1.7 The OBLIST ...........................................   7

2. EXTENSIONS AND DIFFERENCES FROM STANDARD LISP .........   8
2.1 Boolean Arguments And Predicates .....................   8
2.2 Error Handling .......................................   8
2.3 I/O ..................................................  11
2.4 The Interpreter ......................................  15
2.5 The Fap Loader .......................................  18
2.6 Logical Routines And LAP Support .....................  18
2.7 The Alternative Scanner SCAN .........................  20
2.7.1 Scanner Functions ..................................  20
2.8 Other Functions ......................................  22
2.9 Memory Allocation ....................................  24
2.9.1 Storage Allocation Areas ...........................  24
2.9.2 Allocation .........................................  25
2.9.3 Reallocating Space .................................  25
2.10 The Garbage Collector ...............................  26

3. GLOBAL (FLUID) SYSTEM VARIABLES .......................  28

4. ERROR MESSAGES AND WARNINGS ...........................  32
4.1 Error messages not starting with a bad part ..........  32
4.2 Error messages starting with a bad part ..............  36
4.3 Warnings and diagnostic messages .....................  38

5.0 DIFFERENCES BETWEEN LISP 1.6 AND STANDARD LISP .......  40

6.0 REFERENCES ...........................................  42

INDEX ....................................................  43
STANDARD LISP ON DECSYSTEM 10 AND 20                           Page 1


1.0  PRELIMINARIES

This manual describes Standard Lisp implemented on the  DECsystem  10
and  20.  The code for Stanford Lisp 1.6 has been changed in order to
make it conform to the specifications in  the  Standard  Lisp  report
[1].   For this reason the parts of the STANFORD LISP 1.6 MANUAL [2],
that still are relevant have been included here, so that  [2]  is  no
longer  necessary.   This  Standard  Lisp  system contains the system
described in [1] as a subset.  This manual is therefore intended as a
complement  to  [1]  and  features  fully  discussed  in  [1] are not
described here.



1.1  Document Conventions

In the description of the primitive data  structures,  the  following
pictorial notation is used:

        ---------
        |   |   | ->
        ---------
          |
          v

   represents a 36-bit word in FREE STORAGE  (a  Lisp  cell)  with  2
   18-bit  pointers.  The CAR part is to the left and the CDR part to
   the right.


        -----------
        | ABC |   | ->
        -----------

   is similar to the previous, but the ABC stands for  a  pointer  to
   the identifier ABC.  I.e. CAR of the structure is ABC.


        ---------
        |       |
        ---------

   represents a 36-bit word in full word space.


Character conventions:   The  characters  altmode,  carriage  return,
linefeed,  formfeed,  horizontal  tab, space and rubout (delete) will
sometimes be represented as <alt>, <cr>, <lf>, <ff>, <tab>, <sp>  and
<del>.   Control  characters  are usually written as ^ followed by an
ordinary character, e.g. control-A is written ^A.
STANDARD LISP ON DECSYSTEM 10 AND 20                           Page 2


1.2  Special Terminal Control Characters

The time-sharing system treats many  control  characters  in  special
ways.   For  a  complete  discussion  of  control  characters see the
appropriate DECsystem  10  or  20  manuals.   Briefly  the  following
special control characters are used in Lisp.


     Character    Meaning

     ^C           Exit Lisp and talk to the monitor.

     ^G (Bell)    Stop the current Lisp process and  go  to  the  top
                  level  of Lisp.  Only effective when Lisp is asking
                  for input.

     ^O           Suppress  console  printout  until  an   input   is
                  requested.

     ^U           (^X in Tenex) Delete  the  entire  input  line  now
                  being typed.  This is only effective when *DDTIN is
                  NIL.

     <del>        (^A in Tenex) Delete the last  character  typed  if
                  *DDTIN  is  NIL.  If *DDTIN is true, then <del> (in
                  Tenex too)  will  delete  the  entire  S-expression
                  being read.

     ^R           Retypes the current input line.  No special meaning
                  in Tenex.



1.3  Data Types

This section describes the implementation of the primitive data types
discussed  in  [1] 2.1.   It should be noted that some of the details
given in this section  are  temporary  and  may  be  changed.   These
changes  will  be transparent to the user if he does not access parts
of an atom by using CAR, CDR, RPLACA or RPLACD etc..

All atoms (except INUMs, see below) start with one word that contains
a  type tag in the left half and a pointer to the rest of the atom in
the right half.

Integer - There are three representations for integers  depending  on
_______
   the  numerical magnitude of the integer:  INUM, FIXNUM and BIGNUM.
   Their ranges are as follows:

        INUM      abs(n) < K           K is usually 2^16-8

        FIXNUM    K =< abs(n) < 2^35

        BIGNUM    2^35 =< abs(n)
STANDARD LISP ON DECSYSTEM 10 AND 20                           Page 3


   INUMs are small integers represented by pointers  outside  of  the
      normal LISP addressing space.  INUMs are addresses in the range
      2^18-2K-J+1  to  2^18-1-J,  where  J  at  present  is  8.   The
      representation for zero is INUM0 = 2^18-K-J.

   FIXNUMs are represented by list structure of the following form:

           --------------    ---------
           | fixtag |   | -> | value |
           --------------    ---------

      where value is the 2's complement representation of  the  fixed
      point number and fixtag at present is -6.

   BIGNUMs are represented by list structure of the following forms:

           Positive BIGNUMs:

           --------------    ---------           -----------
           | postag |   | -> |   |   | -> ... -> |   | NIL |
           --------------    ---------           -----------
                               |                   |
                               v                   v
                             ---------           -----------
                             | N0     |          | Nn      |
                             ---------           -----------


           Negative BIGNUMs:

           --------------    ---------           -----------
           | negtag |   | -> |   |   | -> ... -> |   | NIL |
           --------------    ---------           -----------
                               |                   |
                               v                   v
                             ---------           -----------
                             | N0     |          | Nn      |
                             ---------           -----------

      where N0 ...  Nn are positive  36  bit  integers  ordered  from
      least  to  most  significant  and  at  present postag is -7 and
      negtag is -8.  The value of a BIGNUM is:
              sign * (N0 + N1*2^35 + N2*(2^35)^2 +...+Nn*(2^35)^n)

   The syntax for integers is  as  described  in  [1] 2.1,  with  the
   exception that an integer followed by a ".", is read as an integer
   with decimal radix instead of a  floating  point  number  with  no
   exponent  and  no  fractional part.  With this exception the radix
   when reading a number is specified by the value  of  the  variable
   IBASE.   The  radix  when  printing an integer is specified by the
   value of the  variable  BASE.   When  the  value  of  BASE  is  10
   (decimal),  then  integers will print with a following ".", unless
   the variable *NOPOINT is not NIL.
STANDARD LISP ON DECSYSTEM 10 AND 20                           Page 4


Floating - Floating point numbers are represented by  list  structure
________
   of the following form:

        --------------    ---------
        | flotag |   | -> | value |
        --------------    ---------

   where value is in DECsystem 10 2's complement  representation  and
   flotag at present is -5.
   The radix for a  floating  point  number  is  always  decimal.   A
   floating point number must be in the (approximate) range:
           10^-38 < abs(x) < 10^+38 or x = 0
   A floating point number has approximately eight significant digits
   of accuracy.

String - Strings are represented by a header word followed by a  list
______
   of full words, each containing five ASCII characters.
   Example:  The string "Time of day" is represented as follows:

        --------------    ---------    ---------    -----------
        | strtag |   | -> |   |   | -> |   |   | -> |   | NIL |
        --------------    ---------    ---------    -----------
                            |            |            |
                            v            v            v
                          ---------    ---------    ---------
                          | Time  |    | of da |    | y^^^^ |
                          ---------    ---------    ---------

   where ^ means null or ASCII 0, and strtag at present is  -4.   The
   number of characters in a string is limited only by available full
   word space.

vector - A vector is internally represented as a dotted pair  of  the
______
   following form:

        --------------
        | vectag |   | -> vector in BPS
        --------------

   where vectag at present is -3.  The  size  of  a  vector  is  only
   limited by the available BPS.

code-pointer - A code pointer is represented as a dotted pair of  the
____________
   following form:

        --------------
        | codtag |   | -> binary code in BPS
        --------------

   where codtag at present is -2.

id - An identifier is represented by a dotted pair of  the  following
__
   form:

STANDARD LISP ON DECSYSTEM 10 AND 20                           Page 5


        -------------
        | idtag |   | -> property list
        -------------

   where idtag at present is -1.
   The syntax for identifiers is similar to that specified in [1] 2.1
   with the following differences:
      <lead-character>     ::= <special-character>|<letter>
      <regular-character>  ::= <lead-character>|<digit>|+|-|'
      <letter>             ::= <alphabetic>|<extra-letter>
      <delimiter>          ::= (|)|<|>|[|]|<alt>|<rub>
      <ignored>            ::= <tab>|<lf>|<vtab>|<ff>|<cr>|<space>|,
      <special-control>    ::= <null>|<control-Z>|'|+|-|.|E|"|!
      <extra-letter>       ::= <any character not a  digit,  ignored,
                                delimiter or special-control>
   The number of characters in an id is  limited  only  by  available
   full word space.



1.4  Property Lists

The property list of an identifier is a list of flags  and  (property
name,  property  value)  pairs  associated with that identifier.  All
identifiers have a PNAME (print name) property with a property  value
that is a list of full words with the same format as for a string.
Example:  The identifier A is represented by:

     ----------    -----------
     | -1 |   | -> |   | NIL |
     ----------    -----------
                     |
                     v
                   -------------    -----------
                   | PNAME |   | -> |   | NIL |
                   -------------    -----------
                                      |
                                      v
                                    ---------
                                    | A^^^^ |
                                    ---------

Example:  After the commands (FLAG '(A) 'NFLG)  and  (PUT 'A 'TYP 'K)
then the identifier A is represented by:

STANDARD LISP ON DECSYSTEM 10 AND 20                           Page 6


   ----------    ---------    ------------    -----------
   | -1 |   | -> |   |   | -> | NFLG |   | -> |   | NIL |
   ----------    ---------    ------------    -----------
                   |                           |
                   v                           v
                 -----------            -------------    -----------
                 | TYP | K |            | PNAME |   | -> |   | NIL |
                 -----------            -------------    -----------
                                                           |
                                                           v
                                                         ---------
                                                         | A^^^^ |
                                                         ---------




1.5  Value Cells

When a value is assigned to an identifier, the property name VALUE is
put  on  the  identifier's  property list with property value being a
pointer to a value cell.  The CDR of the value cell holds  the  value
of  the identifier.  The address of a value cell remains constant for
that identifier unless REMPROPed, to  enable  compiled  functions  to
directly reference the values of global or fluid variables.
Example:   After  the  command  (SETQ A 'B),  the  identifier  A   is
represented by:

     ----------    ---------    -----------
     | -1 |   | -> |   |   | -> |   | NIL |
     ----------    ---------    -----------
                    |              |
                    v              v
             -------------       -------------    -----------
             | VALUE |   | --    | PNAME |   | -> |   | NIL |
             -------------  |    -------------    -----------
                            v                       |
                          ---------                 v
                          | 0 | B |              ---------
                          ---------              | A^^^^ |
                                                 ---------



1.6  Function Cells

When an identifier is defined as a function by PUTD, DE,  DF  or  DM,
the  property  name  FUNCELL is put on the identifier's property list
with property value being a pointer to a function cell.  The  CAR  of
the  function  cell  is the function type, EXPR, FEXPR or MACRO.  The
CDR of the function cell is a lambda expression or  a  code  pointer.
It  is  possible  that  future versions of Standard Lisp will use the
value cell as a function cell.  For that  reason  the  function  cell
should  only  be accessed by means of PUTD, GETD and REMD.  The value
cell should only be accessed by  SET,  SETQ  and  EVAL.   At  present
STANDARD LISP ON DECSYSTEM 10 AND 20                           Page 7


(GETD x) is identical to (GET x 'FUNCELL).
Example:  After the command (PUTD 'A 'type 'body), the  identifier  A
is represented by:

     ----------    ---------    -----------
     | -1 |   | -> |   |   | -> |   | NIL |
     ----------    ---------    -----------
                    |              |
                    v              v
           ---------------       -------------    -----------
           | FUNCELL |   | --    | PNAME |   | -> |   | NIL |
           ---------------  |    -------------    -----------
                            v                       |
                          ---------------           v
                          | type | body |        ---------
                          ---------------        | A^^^^ |
                                                 ---------

where "type" should be one of EXPR, FEXPR or MACRO and "body"  should
be either a lambda expression or a code pointer.



1.7  The OBLIST

For the sake of searching efficiency, the OBLIST has two levels:  the
first level contains sequentially stored "buckets" which are "hashed"
into as a function of the print name of the identifier.  Each  bucket
is  a  list  of  all distinct identifiers which have hashed into that
bucket.  Thus, (CAR OBLIST) is the first bucket, and (CAAR OBLIST) is
the first identifier of the first bucket.
STANDARD LISP ON DECSYSTEM 10 AND 20                           Page 8


2.0  EXTENSIONS AND DIFFERENCES FROM STANDARD LISP

The following functions are implemented exactly  according  to  their
definitions in [1].
ABS, AND, APPEND, ATOM, CONS, DE, DEFLIST,  DELETE,  DF,  DIFFERENCE,
DIVIDE,  DM,  EJECT,  EQ, EQN, EQUAL, EXPAND, EXPT, FIX, FLAG, FLAGP,
FLOAT, FLUIDP, FUNCTION, GENSYM, GET, GETD, GETV, GLOBALP,  GREATERP,
IDP,  INTERN,  LENGTH, LESSP, LIST, LPOSN, MAP, MAPC, MAPCAN, MAPCAR,
MAPCON, MAPLIST, MAX, MAX2, MEMBER, MEMQ, MIN, MINUS,  MIN2,  MKVECT,
NOT,  NULL,  OR,  PAGELENGTH,  PAIR, PLUS, PLUS2, POSN, PRINT, PRIN1,
PRIN2, PROG, PROGN, PUT,  PUTV,  QUOTE,  QUOTIENT,  REMAINDER,  REMD,
REMFLAG,  REMOB,  REMPROP,  REVERSE,  SASSOC,  SUBLIS, SUBST, TERPRI,
TIMES, TIMES2, UPBV.
The rest of this section describes new functions and  functions  that
differ from those in Standard Lisp.



2.1  Boolean Argument And Predicates

In this  Standard  Lisp  implementation  all  booleans  are  actually
extra-boolean,  i.e.  functions  that test boolean arguments test for
NIL or not  NIL,  they  never  test  for  T.   The  predicates  BIGP,
CONSTANTP,  DIGIT,  FILEP, FIXP, FLOATP, INUMP, LITER, NUMBERP, PAIRP
and STRINGP return their  argument  instead  of  T  when  the  tested
condition  is  true.  VECTORP is identical to UPBV so that it returns
the upper limit if its argument is a vector.  Currently CODEP returns
T  if its argument is a code-pointer, but this might change in future
versions.



2.2  Error Handling

Error handling is somewhat different from that described in  [1] 3.8.
Almost  no  Lisp  system  errors  use  the  ERROR function.  They are
handled internally in a similar way as a  call  to  ERROR,  the  only
difference from a call to ERROR, is that the global variable EMSG* is
set to the offending expression (not the entire message) for messages
that include an offending part and to NIL for others.


After an error various types of backtraces will be printed  depending
on  the  value of the variable *BAKGAG.  If *BAKGAG is not NIL then a
backtrace is printed as a series of function calls,  determined  from
the  regular  pushdown  list  (stack),  starting from the most recent
function  call.   If  *BAKGAG  is  0  or  a  negative  integer   then
S-expressions  found  on  the  stack (RPDL) will also be printed.  If
*BAKGAG is +N or -N where N is an integer, then only the N last slots
on   the  stack  will  be  considered  (besides  function  calls  and
S-expressions, the stack may  contain  other  objects  that  are  not
printed).  The format for printing is:
STANDARD LISP ON DECSYSTEM 10 AND 20                           Page 9


     printout            meaning
     ________            _______

     fn1-fn2             Function 1 called function 2

     fn1-EVALARGS        The arguments to  fn1  are  being  evaluated
                         before entering function 1.

     fn1-ENTER           Function 1 is entered.

     ?-fn1               Some internal LISP function called  function
                         1.

     NIL-fn1             Function 1 was called  from  a  place  other
                         than BPS or the Lisp system code.  Something
                         is very wrong.

     =sexp1              The S-expression 1 was on the stack.

Note:   The BACKTRACE printout is often confused by compiled function
        calls  of  the  form  (RETURN (FOO X))  which  is compiled as
        (JCALL (E FOO)) which will not show up in the BACKTRACE.


ERROR(ERVAL:any, MESSAGE:any)
_____________________________
This function is  similar  to  that  in  [1] 3.8  but  ERVAL  is  not
necessarily an integer and the backtrace is somewhat different.

STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 10


EXPR PROCEDURE ERROR(ERVAL,MESSAGE);
 BEGIN SCALAR OCH;
     EMSG!*:=MESSAGE;
     IF NULL !*ERRMSG THEN GOTO ER
      ELSE IF !*ERRMSG NEQ 0 THEN OCH:=WRS NIL;
     IF NULL MESSAGE THEN GOTO BR;
     TERPRI();
     PRIN2 "***** ";
     IF ATOM MESSAGE THEN <<PRIN2 MESSAGE; GOTO BR>>;
     PRIN1 CAR MESSAGE;
     PRINC '! ;
     FOR EACH Y IN CDR MESSAGE DO <<PRIN2 Y; PRINC '! >>;
 BR: IF NOT !*BAKGAG THEN GOTO NB
      ELSE IF NOT FIXP !*BAKGAG THEN
        {backtrace stack back to ERRORSET, printing function calls}
      ELSE IF !*BAKGAG = 0 THEN
        {backtrace stack back to ERRORSET,
           printing function calls and arguments}
      ELSE IF !*BAKGAG > 0 THEN
        {backtrace !*BAKGAG last slots on stack,
           printing function calls}
      ELSE IF !*BAKGAG < 0 THEN
       {backtrace -!*BAKGAG last slots on stack,
           printing function calls and arguments}
     TERPRI();
     IF !*ERRMSG NEQ 0 THEN WRS OCH;
 ER: {return ERVAL to a surrounding ERRORSET, unbinding
       FLUID variables to the environment of the ERRORSET}
 END;


ERRORSET(U:any, !*ERRMSG:any, !*BAKGAG:any):any
_______________________________________________
Type:  EVAL, SPREAD
The difference from the definition in [1] 3.8 are that the second and
third  variables  are  FLUID  and  of  type any.  The effect of these
variables can be seen in the preceding  definition  of  ERROR.   Note
that  messages  are  printed  on  the  standard  output device or the
current output device, not on both.


If the GLOBAL variable !*RSET, of type boolean, is T  then  an  ERROR
that  is  not  surrounded  by ERRORSET (except the top level of Lisp)
doesn't restore the environment of the  top  level.   This  makes  it
possible to examine the variable bindings after an error.  To restore
variable bindings to their top level values, type a bell ( ^G ).


WARNING(MESSAGE:any):NIL
________________________
Type:  EVAL, SPREAD
Prints a message in a format similar to error messages, but  preceded
by  3  asterisks  instead of 5.  If !*ERRMSG is 0 then the message is
printed on the current  output  device,  otherwise  on  the  standard
output device.  If !*MSG is NIL then no message is printed.

STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 11


EXPR PROCEDURE WARNING MESSAGE;
 BEGIN SCALAR OCH;
     IF NOT !*MSG THEN RETURN;
     IF !*ERRMSG NEQ 0 THEN OCH:=WRS NIL;
     TERPRI();
     PRIN2 "*** ";
     IF ATOM MESSAGE THEN <<PRIN2 MESSAGE; GOTO BR>>;
     PRIN1 CAR MESSAGE;
     PRINC '! ;
     FOR EACH Y IN CDR MESSAGE DO <<PRIN2 Y; PRINC '! >>;
 BR: TERPRI();
     IF !*ERRMSG NEQ 0 THEN WRS OCH
 END;


TIME():integer
______________
Type:  EVAL, SPREAD
TIME returns the CPU-time in milliseconds since the  job  was  logged
in.



2.3  I/O

The I/O routines OPEN, CLOSE, RDS and WRS are  similar  to  those  in
[1].   A  description  of  the system dependent parts and the (minor)
differences from [1] is given below.


OPEN(FILENAME-LIST:list, HOW:id):{integer, NIL}
_______________________________________________
OPEN associates the specified files (might be more than one) with one
of  the  16  possible  I/O  channels in Lisp and returns as value the
number of the channel.  This number  is  then  used  as  argument  to
CLOSE, RDS and WRS.  The channels are numbered from 0 to 15.  Channel
0 is the standard device (i.e. the TTY).  It is always open for  both
input and output and cannot be closed.  Instead of 0, NIL can be used
and 0 is always converted to NIL by  RDS  and  WRS  before  assigning
values to the globals INC* and OUTC*.

The syntax for FILENAME-LIST is as follows:

     <filename-list>     ::= (<file-spec>)
                         ::= <id or string>

     <file-spec>         ::= <filename>
                         ::= <file-spec> <filename>
                         ::= <device-name> <file-spec>

     <devicename>        ::= <id or string>:
                         ::= (<atom> <atom>)
                         ::= DIR: <id or string>

     <filename>          ::= <id or string>
                         ::= (<id or string> . <id or string>)
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 12


Semantics:

A filename-list that is an atom is the same as device DSK:, the  atom
as filename and no extension.

A device-name is either an identifier or string ending with colon (:)
which  is the name of an input or output device, or a list containing
a  project-programmer  number,  or  (not  in  TOPS-10  version)  DIR:
followed  by  an  identifier  or  string  which is the name of a disk
directory.  In the two last cases, the disk is implicitly  specified.
I/O  device  names  may  not  have  more than three characters plus a
colon.

A filename is either  an  identifier  or  string  which  specifies  a
filename  with  a  blank  extension, or a dotted pair of filename and
extension.  In both cases the filename applies to  the  recently  (to
the  left)  specified  device-name.   If no device-name is specified,
then DSK: is assumed.  A filename may not be longer than six, and  an
extension not longer than three characters.

HOW must be either INPUT, OUTPUT, INBIN  or  OUTBIN.   The  last  two
keywords  are  used for binary I/O together with BINI and BINO.  They
are mainly intended for the Fap loader.



CLOSE(CHANNELNAME:{integer, NIL}):{integer, NIL}
________________________________________________
Type:  EVAL, SPREAD
CLOSE closes the specified channel if it  is  open  and  returns  its
argument  as  value.  If CHANNELNAME is NIL or 0 then nothing happens
and NIL is returned.



RDS(CHANNELNAME:{integer, NIL}):{integer, NIL}
______________________________________________
Type:  EVAL, SPREAD
RDS selects the specified channel for input and returns as value  the
channelname for the previously selected channel.  When the end of the
selected channel is reached, the  channel  is  automatically  closed.
RDS  sets  the  global  variable  INC*  to  the  channel name for the
currently selected input channel.



WRS(CHANNELNAME:{integer, NIL}):{integer, NIL}
______________________________________________
Type:  EVAL, SPREAD
WRS selects the specified channel for output and returns as value the
channel  name  for  the  previously  selected  channel.  WRS sets the
global variable OUTC* to the channel name for the currently  selected
output channel.
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 13


FILEP(FILENAME-LIST:list):{list, NIL}
_____________________________________
Type:  EVAL, SPREAD
FILENAME-LIST specifies a file in the same way as the first  argument
to  OPEN.   If  this  is  an  existing  file,  then FILEP returns its
argument, otherwise NIL is returned.



SETSYS(SYSDEVICE:any):{integer, string}
_______________________________________
Type:  EVAL, SPREAD
SETSYS redefines the meaning of the device SYS: for OPEN.   It  is  a
different  function  when running on a Tenex machine as compared with
running on a Tops-10 or Tops-20 machine.
On a Tenex machine, if SYSDEVICE is a non-negative integer, then SYS:
is set to mean the directory with directory number SYSDEVICE.
On a Tops machine, if SYSDEVICE is an  identifier  or  string  ending
with  a  colon  ":", then SYSDEVICE is regarded as the name of an I/O
device and SYS: is set to mean that device.  The device name may  not
contain more than three characters, not counting the colon.
If SYSDEVICE is anything else, then nothing is changed.  The value of
SETSYS   is   the  directory  number  respectively  the  device  name
associated with SYS: after the call to SETSYS.



SETPCHAR(PROMPT:{id or string}):id
__________________________________
Type:  EVAL, SPREAD
SETPCHAR sets the Lisp prompter  to  the  four  first  characters  of
PROMPT and returns the previous prompter as a non interned id.



RDSLSH(U:boolean):boolean
_________________________
Type:  EVAL, SPREAD
RDSLSH affects how READ reads the characters "%", "@", "'", "/", "!",
"[", "]", "<" and ">".  If U is NIL then READ is set to read Standard
Lisp, which means that "'" is quote, "%" is start of a comment ending
at  (but not including) <lf>, "!" is escape (slashifier), "[" and "]"
surround a vector, "<" and ">" are super parentheses and "@" and  "/"
are  letters.   If  U  is  not  NIL (e.g. T) then READ is set to read
LISP 1.6, which means that "@" is quote, "/" is escape, "["  and  "]"
are  super  parentheses  and  "%", "<", ">", "'" and "!" are letters.
RDSLSH ends by calling SCANSET(NIL).



PGLINE():pair
_____________
Type:  EVAL, SPREAD
PGLINE returns a dotted pair of current  line  and  current  page  of
input.  When the input is coming from the TTY, then the the answer is
always (1 . 1).  When the input is coming from a  file  with  SOS  or
EDIT  line  numbers,  then those line numbers are returned, otherwise
the line and page numbers are found by counting line and page feeds.
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 14


READ():any
__________
Type:  EVAL, SPREAD
READ is the same as in  [1]  with  the  following  differences.   The
delimiters "<" and ">" operate as "super-parentheses".  A right angle
bracket ">" will close all open left parentheses "(" up  to  matching
left  angle bracket "<".  If there is no matching left angle bracket,
it will close the entire S-expression.
When the input is comming  from  the  terminal,  then  the  character
altmode   (or   escape)  will  end  the  input  line  and  close  the
S-expression being read in the same way  as  a  rigth  angle  bracket
without a matching left angle bracket.



SKIPTO(SKPCH:{id, string}):integer
__________________________________
Type:  EVAL, SPREAD
SKIPTO skips characters from  the  selected  input  channel  until  a
character is found that matches the first character in the print name
of SKPCH.  The ASCII code for that character is returned as value.



TYI():integer
_____________
Type:  EVAL, SPREAD
TYI reads the next character from  the  selected  input  channel  and
returns the ASCII code for that character.



UNREADCH(BCHR:id):id
____________________
Type:  EVAL, SPREAD
UNREADCH puts the first character in BCHRs print name in an  internal
backspace  buffer  and  returns BCHR.  This has as effect that when a
character is to be read from the selected input channel, then  it  is
instead  taken from the backspace buffer and the buffer is cleared so
that subsequent characters are read from the selected input  channel.
The  buffer  can  only  contain one character at a time.  If a second
UNREADCH is executed before the buffer is cleared, then the buffer is
overwritten  with  the  new  BCHR,  undoing  the  effect of the first
UNREADCH.
UNREADCH should only be used after TYI or  READCH.   The  reason  for
this  is  that  functions  that  read atoms (READ and SCAN), uses the
backspace buffer to store the delimiter.



TYO(OCHR:integer):integer
_________________________
TYO prints the character whose ASCII code is OCHR and returns OCHR.



PAGELENGTH and LINELENGTH
__________     __________
are similar to their definition in [1] but there is no upper limit to
their  argument.   The  argument  should be a positive number or NIL.
PAGELENGTH also accepts  0  as  argument.   Calls  to  PAGELENGTH  or
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 15


LINELENGTH  do  only  affect  the  presently selected output channel.
Each output channel that is  OPENed,  has  initially  LINELENGTH  112
(decimal)  and  PAGELENGTH  0  (i.e.  no  paging).   The  TTY channel
(channel NIL) has initially LINELENGTH 69 and PAGELENGTH 0.



The character ^Z (Ascii 32) is used by  PRIN1,  PRIN2  and  PRINT  to
break  an  atom  that  is to long to fit in on the current line.  All
reading routines  (TYI,  READCH  and  READ)  will  ignore  everything
between a ^Z and a <lf> including the end points.



Standard Lisp uses the character ^_ (Ascii 37) internaly  as  end  of
line character.  The reading routines will interpret a single <lf>, a
single <ff> or a <cr> followed by a <lf> or a <ff> as an end of line.
The printing routines print ^_ as <cr><lf>, except for TYO:  (TYO 37)
(octal) prints ^_ (control ).



2.4  The Interpreter

The interpreter is an extension of that defined  in  [1].   The  main
differences from [1] are:
     The body of a lambda expression is an implied PROGN.
     Environments can be passed around (FUNARG feature).
     The LABEL construct is supported.
     FEXPRs may have a second argument that is the environment.
The compiler doesn't support these features.



EVAL(U:any):any
_______________
Type:  EVAL, SPREAD

STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 16


EXPR PROCEDURE EVAL U;
 BEGIN SCALAR FN;
  RETURN
   IF ATOM U THEN
     IF NOT IDP U THEN U
      ELSE IF A:=GET(U,VALUE) AND NOT(A:=CDR A EQ 'UNBOUND) THEN A
      ELSE ERROR(29,LIST(U,"UNBOUND VARIABLE))
    ELSE IF IDP CAR U THEN
     IF A:=GETD CAR U THEN
      IF CAR A EQ 'EXPR THEN APPLY(CDR A,EVLIS CDR U)
       ELSE IF CAR A EQ 'FEXPR THEN APPLY(CDR A, LIST CDR U)
       ELSE IF CAR A EQ 'MACRO THEN EVAL APPLY(CDR A, LIST U)
       ELSE ERROR(28,LIST(CAR U,"UNDEFINED FUNCTION"))
      ELSE IF A:=GET(CAR U,'VALUE) AND NOT((A:=CDR A) EQ 'UNBOUND)
        AND NOT(A EQ CAR U) THEN EVAL A . CDR U
      ELSE ERROR(28,LIST(CAR U,"UNDEFINED FUNCTION"))
     ELSE IF PAIRP CAR U OR CODEP CAR U
      THEN APPLY(CAR U, EVLIS CDR U)
     ELSE ERROR(28,LIST(CAR U,"UNDEFINED FUNCTION"))
 END;



Note:   The identifier UNBOUND is not interned and the ERROR function
        is not used.



APPLY(FN:any, ARGS:list):any
____________________________
Type:  EVAL, SPREAD

EXPR PROCEDURE APPLY(FN,ARGS);
 BEGIN SCALAR A;
  RETURN
   IF ATOM FN THEN
     IF IDP FN THEN
       IF A:=GETD FN THEN
         IF CAR A EQ 'EXPR THEN APPLY(CDR A, ARGS)
          ELSE ERROR(31,LIST(FN,"NOT EXPR - APPLY"))
        ELSE IF A:=GET(FN,'VALUE) AND NOT((A:=CDR A) EQ 'UNBOUND)
          AND NOT(A EQ FN) THEN APPLY(A,ARGS)
        ELSE ERROR(30,LIST(FN,"UNDEFINED FUNCTION - APPLY"))
      ELSE IF CODEP FN THEN
       IF LENGTH ARGS > 14 THEN ERROR(145,"TO MANY ARGS FOR EXPR")
        ELSE {spread the actual parameters in ARGS, transfer to entry
              point of the function and return the value returned
              by the function}
      ELSE ERROR(30,LIST(FN,"UNDEFINED FUNCTION - APPLY"))
    ELSE IF CAR FN EQ 'LAMBDA THEN
     IF A:=CADR FN AND IDP FN THEN ERROR(146,"ILLEGAL LAMBDA FORMAT")
      ELSE IF (A:=LENGTH A - LENGTH ARGS) < 0 THEN
        ERROR(147,"TO MANY ARGUMENTS SUPPLIED - APPLY")
      ELSE IF A > 0 THEN
       IF A=1 AND LENGTH ARGS =1 AND
           {APPLY was called from EVAL to evaluate FEXPR call} THEN
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 17


          APPLY(FN,LIST(CAR ARGS,{binding context pointer (BCP)}))
        ELSE ERROR(148,"TO FEW ARGUMENTS SUPPLIED - APPLY")
      ELSE <<{bind the actual parameters in ARGS to the formal
               parameters in CADR FN};
             A:=EVAL('PROGN . CDDR FN);
             {unbind formal parameters};
             A>>
    ELSE IF CAR FN EQ 'FUNARG THEN
     <<{unbind parameters back to BCP in CADDR FN};
       A:=APPLY(CADR FN,ARGS);
       {restore bindings};
       A>>
    ELSE IF CAR FN EQ 'LABEL THEN
     <<{bind function in CADDR FN to formal name in CADR FN};
       A:=APPLY(CADDR FN,ARGS);
       {unbind name};
       A>>
    ELSE APPLY(EVAL FN,ARGS)
 END;



!%EVAL(U:any,ENV:{list, integer}):any
_____________________________________
Type:  EVAL, SPREAD

EXPR PROCEDURE !%EVAL(U,ENV);
 BEGIN SCALAR A;
  IF PAIRP ENV THEN FOR EACH X IN ENV DO
     {bind actual parameter in CDR X to formal parameter in CAR X}
   ELSE {ENV is a BCP. Unbind parameters back to BCP};
  A:=EVAL U;
  {restore parameters};
  RETURN A
 END;



!%APPLY(FN:any,ARGS:list,ENV:{list, integer}):any
_________________________________________________
Type:  EVAL, SPREAD

EXPR PROCEDURE !%APPLY(FN,ARGS,ENV);
 !%EVAL(LIST('APPLY,FN,ARGS),ENV);



!*FUNCTION(FN:any)
__________________
Type:  EVAL, SPREAD
!*FUNCTION is used instead of FUNCTION when it is necessary  to  pass
along an environment together with a function definition.

FEXPR PROCEDURE !*FUNCTION FN;
 LIST('FUNARG,CAR FN,{BCP for current environment});
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 18


2.5  The Fap Loader

The Fap loader loads binary files generated by the compiler.



FLOAD(FILE-LIST:list):NIL
_________________________
Type:  NOEVAL, NOSPREAD
FILE-LIST is a list of identifiers or strings, each of which  is  the
name  of  a  file  with the implied extension FAP.  FLOAD loads these
files with the Fap loader.

FEXPR PROCEDURE FLOAD FILES;
 FOR EACH X IN FILES DO FISLM X;



FISLM(FILE:id):NIL
__________________
Type:  EVAL, SPREAD
FISLM loads the Fap file FILE.FAP.



If the global variable *PREDEF is not NIL then the Fap loader  prints
the message:
        *** x REDEFINED
if the function x is redefined.

If the global variable *PURIFY is not NIL then the  Fap  loader  will
try  to  convert  the  function links in the loaded binary code, from
slow to fast links.



2.6  Logical Routines And LAP Support

The following routines are  mainly  intended  for  use  by  the  Lisp
Assembler (LAP).



!*BOX(POINTER:any):integer
__________________________
Type:  EVAL, SPREAD
!*BOX converts POINTER from its  machine  representation  to  a  Lisp
number.  !*BOX is the inverse operation to NUMVAL.



NUMVAL(NUM:number):any
______________________
Type:  EVAL, SPREAD
NUMVAL converts the number NUM from its Lisp  representation  to  its
machine  representation.   NUMVAL(!*BOX(x))  is  EQ  to  x for any x.
!*BOX(NUMVAL(x)) is EQUAL to x if x is an INUM or a FIXNUM.
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 19


MKCODE(ADR:integer, NARG:integer):codepointer
_____________________________________________
Type:  EVAL, SPREAD
MKCODE converts the address ADR into a code  pointer.   NARG  is  the
number of arguments.



BOOLE(OPR:integer, A:integer, B:integer):integer
________________________________________________
Type:  EVAL, SPREAD
BOOLE performs a 36 bit Boolean operation on A and B.  OPR  specifies
which  of  16  Boolean  operations  to perform.  These operations are
defined in the following table of the result for the i:th bit in  the
answer with given value of OPR and i:th bit of A and B.  In the table
"-" stands for boolean not.
              A,B        0,0       1,0       0,1       1,1
     OPR
       0       0          0         0         0         0
       1      A^B         0         0         0         1
       2     -A^B         0         0         1         0
       3       B          0         0         1         1
       4      A^-B        0         1         0         0
       5       A          0         1         0         1
       6     -A=B         0         1         1         0
       7      AvB         0         1         1         1
       8     -A^-B        1         0         0         0
       9      A=B         1         0         0         1
      10      -A          1         0         1         0
      11     -AvB         1         0         1         1
      12      -B          1         1         0         0
      13      Av-B        1         1         0         1
      14     -Av-B        1         1         1         0
      15       1          1         1         1         1

As  can  be  seen  from  above,  OPR  is  the  number  whose   binary
representation is the corresponding table entry.



LSH(NUM:integer, SHIFT:integer):integer
_______________________________________
Type:  EVAL, SPREAD
LSH performs a logical 36 bit left shift of SHIFT places on NUM.   If
SHIFT  is  negative,  then NUM will be shifted right.  In both cases,
vacated bits are filled with zeros.



EXAMINE(ADR:integer):integer
____________________________
Type:  EVAL, SPREAD
EXAMINE converts the word at address ADR in memory to a  Lisp  number
and returns it.
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 20


DEPOSIT(NUM:number, ADR:integer):integer
________________________________________
Type:  EVAL, SPREAD
DEPOSIT deposits the 36 bit machine representation of NUM in the word
at address ADR in memory.  NUM is returned as value.



2.7  The Alternative Scanner SCAN

Lisp uses a table driven scanner.  There are two scanner tables,  the
Standard  Lisp  table  and a second table that may be modified by the
user for the purpose of implementing scanners for other languages.
The syntax when using the second table is:
     <alphabetic>   ::= A|B| ...  |Z|a| ...  |z
     <extra letter> ::= <initially no character>
     <letter>       ::= <alphabetic>|<extra letter>
     <ignored>      ::= <tab>|<lf>|<vtab>|<ff>|<cr>|<space>
     <digit>        ::= 1|2|3|4|5|6|7|8|9
     <delimiter>    ::= <all  characters  except  letter,  digit   or
                        ignored>
     <special control> ::= <null>|^Z|+|-|.|E|<comment start>|<comment
                        end>|<string start>|<string end>|<slashifier>
     <identifier>   ::= <letter>
                    ::= <slashifier><any character>
                    ::= <identifier><digit>
                    ::= <identifier><identifier>
     <comment>      ::= <comment start><any characters except comment
                        end>
                    ::= ^Z<any characters except lf or ff><lf or ff>
     <string-text>  ::= <any character except string end>
                    ::= <string end><string end>
                    ::= <string-text><string-text>
     <string>       ::= <string start><string-text><string end>
    The syntax for numbers is as usual (see 1.3), except that leading
    signs ("+" or "-") in numbers are treated as delimiters.

In the  above  syntax  letter,  delimiter,  ignored,  comment  start,
comment  end,  string start, string end and slashifier can be changed
by the user.  Initially comment start is "%", comment  end  is  <lf>,
string start and string end is " (double quote) and the slashifier is
"!", i.e. the same as for the ordinary Lisp scanner.  Note that  "^Z"
starts  a  comment  also  when  comment  start  has  been  defined as
something else.  Note also that the comment end is not commented  out
in a comment started by something else than ^Z.



2.7.1  Scanner Functions - In   the   following   descriptions,   all
characters are specified by their ASCII code.  E.g. " " is (in octal)
40 and "A" is 101.


SCANSET(SWITCH:extraboolean):extraboolean
_________________________________________
Type EVAL, SPREAD
SCANSET switches the  scanners  attention  between  the  two  scanner
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 21


tables.   If  SWITCH  is NIL the Standard Lisp scanner table is used,
otherwise the user settable scanner table is used.   SCANSET  returns
the previous setting.


SCANINIT(COMMENT-START:integer, COMMENT-END:integer,
____________________________________________________
   STRING-START:integer, STRING-END:integer, SLASHIFIER:integer):NIL
   _________________________________________________________________
Type EVAL, SPREAD
SCANINIT reinitializes the scanner  so  that  all  characters  except
alphabetics  and  digits,  are delimiters (no ignored).  It then sets
comment start, comment end, string start, string end  and  slashifier
to be the specified characters.


DELIMIT(CHAR:integer):integer
_____________________________
Type EVAL, SPREAD
DELIMIT specifies to the scanner that CHAR is a delimiter.   CHAR  is
returned as value.


LETTER(CHAR:integer):integer
____________________________
Type EVAL, SPREAD
LETTER specifies to the scanner that  CHAR  is  a  letter,  and  thus
allows CHAR to be in an identifier.  CHAR is returned as value.


IGNORE(CHAR:integer):integer
____________________________
Type EVAL, SPREAD
IGNORE specifies to the scanner that CHAR is not to be returned as  a
delimiter from scan, but instead will be ignored.  However, CHAR will
still function as a delimiter inside a number or an identifier.


SCAN():integer
______________
Type EVAL, SPREAD
SCAN reads an atom or delimiter and sets the variable SCNVAL  to  the
value  read,  and  returns  a number corresponding to the syntactical
type read as follows:
     Type           SCAN value     SCNVAL value

     identifier     0              the uninterned identifier
     string         1              the string
     number         2              the number
     delimiter      3              an interned id with the  delimiter
                                   as print name


SCAN will give different results depending on which scanner is  used.
READ  checks which scanner is used and will temporarily switch to the
Standard Lisp scanner if that scanner is not already used.  An  error
in  READ  might leave the Standard Lisp scanner in use even though it
wasn't before the READ.
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 22


2.8  Other Functions

INTERNP(U:any):extra-boolean
____________________________
Type:  EVAL, SPREAD
INTERNP returns T if U is an interned id, otherwise NIL.


INUMP(U:any):extra-boolean
__________________________
Type:  EVAL, SPREAD
INUMP returns U if it is an INUM, otherwise NIL.


BIGP(U:any):extra-boolean
_________________________
Type:  EVAL, SPREAD
BIGP returns U if it is a BIGNUM, otherwise NIL.


GCD(A:integer, B:integer):integer
_________________________________
Type:  EVAL, SPREAD
GCD returns the greatest common divisor of A and B.


COMPRESS(U:id-list):any
_______________________
Type:  EVAL, SPREAD
COMPRESS is similar to the definition in [1],  but  it  is  currently
more general.  It uses READ to read the character ids in U and it can
read an arbitrary S-expression.


EXPLODE(U:any):id-list
______________________
Type:  EVAL, SPREAD
EXPLODE is similar to the definition in [1], but it is currently more
general  as  it  can explode any S-expression.  This is done by using
PRIN1 to print to a list.  EXPLODE uses  the  Standard  Lisp  scanner
table and will temporarily switch scanner tables if necessary.


EXPLODEC(U:any):id-list
_______________________
Type:  EVAL, SPREAD
EXPLODEC is similar to EXPLODE but it uses PRIN2  instead  of  PRIN1.
This  means  that  the  escape  character  does  not  prefix  special
characters and strings are not enclosed by " ... ".


SET and SETQ
___     ____
are the same as in [1]  except  that  they  do  not  declare  unknown
variables to be FLUID.


FLUID, UNFLUID and GLOBAL
______ _______ ___ ______
are similar to their definition in [1] but they presently  have  some
side effects used by REDUCE.
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 23


PUTD(U:id,V:id,W:any):id
________________________
Type:  EVAL, SPREAD.
If U is flagged LOSE, then PUTD does nothing.  If V is EXPR, FEXPR or
MACRO,  then  U  is  defined  as  a  function of type V with function
property W else PUTD is the same as PUT.
If *COMP is on and FLAGP(V,'COMPILE) and W  is  a  LAMBDA  expression
then W is compiled.  Initially EXPR and FEXPR are flagged COMPILE.


ATSOC(U:any, V:alist):{dotted-pair, NIL}
________________________________________
Type:  EVAL, SPREAD
ATSOC is similar to ASSOC but it uses EQ instead of EQUAL.

EXPR PROCEDURE ATSOC(U, V);
 IF NULL V THEN NIL
  ELSE IF U EQ CAAR V THEN CAR V
  ELSE ASSOC(U, CDR V);


ASCII(COD:integer):integer
__________________________
Type:  EVAL, SPREAD
ASCII returns an non-interned single character identifier  such  that
the ASCII code for the character is COD.
Example:   (ASCII 101) returns an identifier with print name "A".


COND([U:cond-form]):any
_______________________
Type:  NOEVAL, NOSPREAD
Cond-form is more general than that defined in [1] 2.3.  A  cond-form
is a list of N+1 element lists of the form:
        (ANTECEDENT:any CONSEQUENT-1:any ...  CONSEQUENT-N:any)
where N is a non-negative integer.
The Antecedents are evaluated in their order of  appearance  until  a
non-NIL  value is encountered.  The consequents of the selected U are
evaluated from left to right and the value of the last consequent  is
returned  as  value  of  COND.   If there are no consequents then the
value of the antecedent is returned.


FREEZE(U:boolean)
_________________
Type:  EVAL, SPREAD
FREEZE halts the Lisp system in such a  way  that  it  can  later  be
restarted.  The value of U determines what happens when the system is
restarted:  If U is true (not NIL)  then  any  additional  memory  is
allocated  and  the the system is back at the top level.  If U is NIL
then the system is restored to where it  was  (except  for  that  all
files are closed), and NIL is returned as value of FREEZE.
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 24


2.9  Memory Allocation

The Standard Lisp system has  many  different  areas  of  memory  for
storing  data  which  can  independently  vary  in  size.   Some LISP
applications demand larger allocations for these areas  than  others.
To allow users to adjust the sizes of these areas to their own needs,
a memory allocation procedure exists.



2.9.1  Storage Allocation Areas -

     BINARY PROGRAM SPACE     Compiled functions and vectors.
     FREE STORAGE             LISP pairs
     FULL WORD SPACE          Numbers and print names
     BIT TABLES               Used by garbage collector
     REGULAR PUSHDOWN LIST    Return addresses for function calls and
                              local   variables  in  interpreter  and
                              compiled code.
     SPECIAL PUSHDOWN LIST    Old bindings of fluid variables.
     EXPANDED CORE            I/O buffers.

The various areas are differently allocated on a Tops-10  machine  as
compared with a Tops-20 or Tenex machine.  The memory map for the two
cases are as follows:

            TOPS-10                         TENEX or TOPS-20

                                               TOP OF CORE
                                        -------------------------
                                        | BINARY PROGRAM SPACE  |
                                        -------------------------
                                           BOTTOM OF HIGH CORE

            TOP OF CORE                      TOP OF LOW CORE
     -------------------------          -------------------------
     |     EXPANDED CORE     |          |     EXPANDED CORE     |
     -------------------------          -------------------------
     | SPECIAL PUSHDOWN LIST |          | SPECIAL PUSHDOWN LIST |
     -------------------------          -------------------------
     | REGULAR PUSHDOWN LIST |          | REGULAR PUSHDOWN LIST |
     -------------------------          -------------------------
     |      BIT TABLES       |          |      BIT TABLES       |
     -------------------------          -------------------------
     |    FULL WORD SPACE    |          |    FULL WORD SPACE    |
     -------------------------          -------------------------
     |     FREE STORAGE      |          |     FREE STORAGE      |
     -------------------------          -------------------------
     | BINARY PROGRAM SPACE  |          |   LISP INTERPRETER    |
     -------------------------          -------------------------
     |   LISP INTERPRETER    |                 BOTTOM OF CORE
     -------------------------
            BOTTOM OF CORE
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 25


2.9.2  Allocation -
When the Lisp system is initially started, it asks the user:
               ALLOC?
If this is answered with "Y" then the following questions are  asked,
otherwise  the  default values are used.  The questions vary slightly
between a Tops-10, a Tops-20 and a Tenex machine:
Tops-10  Tops-20    Tenex      Default       Meaning

         Core (K):  Core (K):  12            K words core (decimal).

SYS:     SYS:                  SYS:          Meaning of SYS: in OPEN.
                    SYS: dir#  Dir# for      Meaning of SYS: in OPEN.
                               <REDUCE> or 0.

FWDS=    FWDS=      FWDS=      400+1/4FREE   Words  full  word  space
                                             (octal)
BPS.=                          2000          Words   binary   program
                                             space (octal)

SPDL=    SPDL=      SPDL=      1000          Words  special  pushdown
                                             list (octal)
RPDL=    RPDL=      RPDL=      1000          Words  regular  pushdown
                                             list (octal)
HASH=    HASH=      HASH=      77            Number  of  buckets   in
                                             oblist (octal)

There are two basic responses to any of these questions:
     1.  A space causes the default value to be used.
     2.  A number ended by a space or a cr causes the  number  to  be
         used.




2.9.3  Reallocating Space -



CORE(SIZE:integer):integer
__________________________
Type:  EVAL, SPREAD
If SIZE is 0 then CORE returns the (low-)core size including extended
core.   If  SIZE  is  negative  then CORE returns the (low-)core size
excluding extended core (i.e.  Lisp  allocated  core).If  SIZE  is  a
positive  integer  then  the system will try to allocate SIZE K words
(low-)core.
   If SIZE is smaller than current Lisp alloc. core, then the warning
   message:
   *** CAN'T CUT CORE INTO ALLOCATED SPACE
   is printed, no action is taken and the value returned is the  same
   as  that  when  SIZE  is  negative (i.e. minimum positive value of
   SIZE).
   Otherwise the core will be expanded to SIZE K words and (if not  a
   (EXCORE  n)  has  been  executed) then the additional core will be
   reallocated in the following way:
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 26


        1/4 for full words space
        1/64 for each push down list
        about 1/16 for bit tables
        the remainder for free storage (Lisp cells)



The above described allocation scheme is also used if the run of Lisp
is  stopped (by an error or by a ^C) and then either started again by
the START command or SAVEd and then RUN again.  When BPS is  in  high
core  (Tops-20 or Tenex) then it is automatically expanded as long as
there is free high core.  When BPS is in low core (Tops-10) then  The
binary  program  space  can  not  actually  be expanded, but an extra
binary program space can be defined  in  the  expanded  core  by  the
following function.



EXCORE(ARG:any):any
___________________
Type:  EVAL, SPREAD
EXCORE has two effects:
     1.  The allocation procedure described in CORE, is inhibited  if
         ARG is not NIL.
     2.  In case of a Tops-10 machine, there is also a second effect:
         If ARG is a positive integer then  ARG  K  words  of  binary
         space  is  defined in extra core and BPORG and BPEND are set
         to the beginning and end of this extra BPS.  If ARG is 0  or
         NIL  then  BPORG  and  BPEND  are set to their old values in
         ordinary BPS if they have been set to point into extra BPS.




EXCISE():T
__________
Type:  EVAL, SPREAD
EXCISE contracts core to its size excluding extra core.
Note that if a Lisp core image is saved and later RUN, then the extra
core will be reallocated as described in CORE.



2.10  The Garbage Collector

The garbage collector in LISP is called whenever FREE STORAGE or FULL
WORD  SPACE  is exhausted.  The garbage collector analyzes the entire
state of list structure which is pointed to by either the OBLIST, the
regular  pushdown list, the special pushdown list, list arrays, and a
few other special cells.  By recursively marking all  words  in  free
and  full  word  spaces  which  are  pointed to in this manner, it is
possible to  determine  which  words  are  not  pointed  to  and  are
therefore  garbage.   Such  words  are  collected  together  on their
respective free storage or full word space lists.
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 27


RECLAIM():NIL
_____________
Type:  EVAL, SPREAD
RECLAIM causes a garbage collection to occur and returns NIL.



GCTIME():integer
________________
Type:  EVAL, SPREAD
GCTIME returns the number of  milliseconds  LISP  has  spent  garbage
collecting since it was started.



SPEAK():integer
_______________
Type:  EVAL, SPREAD
SPEAK returns the total number of CONSes  which  have  been  executed
since LISP was started.
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 28


3.0  GLOBAL (FLUID) SYSTEM VARIABLES

        *RAISE, NIL and T are implemented as described in [1]..



        The following variables differ from those defined in [1].



*COMP - Initial value = NIL
_____
See description of PUTD in section 2.8.



EMSG* - Initial value = NIL
_____
Most of the system errors do not use ERROR and  only  part  of  their
message is found as the value of EMSG*.  See ERROR.



$EOF$ - Initial value = $EOF$
_____
$EOF$ is returned to a surrounding ERRORSET  (or  the  top  level  if
there is none) when a input function reaches the end of the currently
selected input file.



$EOL$ - Initial value = ^_ (Ascii 37)
_____
See description of ^_ at the end of section 2.3.



        The following variables are extensions of Standard Lisp:



*BAKGAG - Initial value = NIL
_______
*BAKGAG is the third argument to ERRORSET.  If  *BAKGAG  is  not  NIL
then  a  backtrace  will  be  printed  when an error occurs.  For the
different forms of backtraces depending on the value of  *BAKGAG  see
ERROR.



*DDTIN - Initial value = NIL
______
*DDTIN controls whether input from the TTY is seen line  by  line  or
character by character.
If *DDTIN is NIL, then input is seen line by  line,  i.e.  the  input
buffer is only sent to the Lisp system when a carriage return, a line
feed, a form feed or an altmode is typed on the TTY.  This means that
terminal  editing characters like ^U, <rub> etc. (or corresponding in
a Tenex system) can be used.
If *DDTIN is not NIL  (e.g.  T)  then  input  is  seen  character  by
character,  i.e. each character is sent to the Lisp system when it is
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 29


typed on the TTY.  This means that editing characters can't be  used,
but  <rub>  will  delete the entire S-expression being read and start
reading again.



*ERRMSG - Initial value = T
_______
*ERRMSG is the second argument to ERRORSET.  If *ERRMSG is  NIL  then
no  error  message  will  be  printed.   If  *ERRMSG  is 0 then error
messages will be printed on the currently  selected  output  channel.
If  *ERRMSG  is anything else, then error messages will be printed on
the TTY.



*ECHO - Initial value = NIL
_____
If *ECHO is not NIL then all characters read from a  file  (not  TTY)
are  echoed  to the current output file.  No conversion from lower to
upper case is done of the echoed character regardless of the value of
!*RAISE.   If  a ^D is encountered in the file, the program will wait
until a " " (a space) is typed on the TTY and then continue.  The  ^D
is not echoed.
If *ECHO is NIL no echoing takes place.



*GCGAG - Initial value = NIL
______
If *GCGAG is not NIL  when  a  garbage  collection  occurs  then  the
following is printed on the TTY:
either
        *** FREE STG EXHAUSTED
or
        *** FULL WORD SPACE EXHAUSTED
or nothing
followed by
        *** n FREE STG, m FULL WORDS AVAILABLE
where n and m are the number of words (cells) printed in the  current
radix.
If *GCGAG is NIL then these messages are not printed.



*MSG - Initial value = T
____
If *MSG is NIL then no warnings will be printed, e.g. PUTD  will  not
print any message when a function is redefined.



*NOPOINT - Initial value = NIL
________
If *NOPOINT is NIL then integers printed (or EXPLODEd) with a decimal
radix (BASE = 10.) are followed by a ".".
If *NOPOINT is not NIL then decimal integers are not printed  with  a
following ".".
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 30


*NOUUO - Initial value = NIL
______
If *NOUUO is not NIL then slow links in compiled code  are  converted
to  fast links when the code is executed.  Fast links can't be TRaced
as can slow links.



*PREDEF - Initial value = NIL
_______
If *PREDEF is not NIL then the Fap loader will print
        *** x REDEFINED
on the TTY, when x is a function that is being redefined.
If *PREDEF is NIL then no message is printed.



*PURIFY - Initial value = NIL
_______
If *PURIFY is not NIL then the Fap loader will try  to  convert  slow
links to fast links in the loaded binary code.
If *PURIFY is NIL then this conversion will not take  place  at  load
time.   It  will  take place when executing the code if *NOUUO is not
NIL.



*RSET - Initial value = NIL
_____
If an error occurs that is not surrounded by an ERRORSET so that  the
system  will  go to the top level, then the value of *RSET determines
whether fluid variables will be restored to their top level values or
kept at the values they had when the error occurred:
If *RSET is NIL then fluid variables will be restored  to  top  level
values.
If *RSET is not NIL then the current values will be kept.



BASE - Initial value = 10 (decimal)
____
The value of BASE is used as radix when printing integers.   Floating
point numbers are always decimal.



IBASE - Initial value = 10 (decimal)
_____
The value of IBASE is used as radix when reading integers.   Floating
point numbers are always decimal.



INITFN* - Initial value = NIL
_______
If INITFN* is not NIL, then it must be a function  of  no  arguments.
This  is  an  initialization  function that will be evaluated after a
Lisp error return to the top level has occurred or when  a  BELL  has
been  typed  or  when  a Lisp core image is started again or when the
memory is reallocated by CORE.
INITFN* is useful when it is desirable to change  the  top  level  of
Lisp.
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 31


          (SETQ INITFN!* (FUNCTION EVALQUOTE))
causes the top level to become EVALQUOTE instead of EVAL.



BPEND - Value = Address of end of BPS
_____
(NUMVAL BPEND) points to the end of BPS.



BPORG - Value = Address of beginning of free BPS
_____
(NUMVAL BPORG) points to the beginning of free BPS.



INC* - Value = Channel number for currently selected  input  channel.
____
INC*  is  updated  by  calls  to RDS and should not be changed in any
other way.  INC* is NIL when the current input channel is the TTY.



OUTC* - Value = Channel number for currently selected output channel.
_____
OUTC*  is  updated  by  calls to WRS and should not be changed in any
other way.  OUTC* is NIL when the current output channel is the TTY.



OBLIST - Value = List of lists of interned identifiers.
______
See 1.7



SCNVAL - Initial value = NIL
______
See SCAN.



SYSTEM* - Value = 1, 0 or -1.
_______
SYSTEM* indicates what operating system the current  Lisp  system  is
intended for:
        System      Value of SYSTEM*

        Tenex            1
        Tops-10          0
        Tops-20         -1
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 32


4.0  ERROR MESSAGES AND WARNINGS

The Lisp system checks for some error conditions and prints  messages
accordingly,  but  some  of  the  error  checking described in [1] is
currently not implemented.
In particular CAR, CDR, RPLACA and RPLACD do  not  check  that  their
argument is a pair.  This means also that improper arguments to NCONC
or ASSOC are not checked.
Illegal use of GO or RETURN is not checked, indeed GO and RETURN  can
currently be used almost anywhere in interpreted code.
The use of an id as both a  function  and  a  variable  is  currently
allowed.
Erroneous conditions (e.g.  illegal CAR or CDR) that are not  tested,
result  in  either  the wrong message at some later time, or no error
message at all.  In the latter case the  system  has  screwed  itself
without complaining.
When error messages are printed, it is usually difficult to determine
which  function  caused  the error and which functions called it.  In
this situation, the BACKTRACE is useful,  (see  *BAKGAG).   *RSET  is
also  useful  to determine the values of variables at the time of the
error.
The following is a list of error messages, their cause  and  in  some
cases their remedy.  The messages are grouped in four groups and they
are listed  alphabetically  within  each  group.   The  "*****"  that
precedes  error  messages  and  the "***" that precedes warnings have
been omitted  in  the  description.   The  error  messages  for  some
optional  features that are not normaly available, are included.  The
descriptions of possible causes for the  errors,  sometimes  includes
functions that are not normaly available.



4.1  Error Messages Not Starting With A Bad Part

These messages are of the type:
     ***** <message>.

BIGNUM UNSUITABLE AS ARGUMENT
        BOOLE, DEPOSIT, EXAMINE, BINO, OPEN, ASCII,  MKCODE,  SETMOD,
        CPLUS,  CDIF,  CTIMES, CRECIP, MKVECT, LSH, %SOSSWAP, WBLK or
        EXCORE, was given a number that was not an INUM or a FIXNUM.
        This error might also occur in some other routines, if  BPORG
        or BPEND have been set to BIGNUMs.

BINARY PROGRAM SPACE EXCEEDED
        LAP, (ARRAY, EXARRAY or LOAD)  has  exceeded  BINARY  PROGRAM
        SPACE.   If  you  are  running on a Tenex or Tops-20 machine,
        this should not happen as the available BPS is about  100  K.
        If  you are running on a Tops-10 machine you can either start
        again from a fresh LISP and allocate more BPS, or you can use
        EXCORE.

CAN'T EXPAND CORE
        CORE, EXCORE, OPEN, (LOAD or ED) failed to expand  core.   In
        the  case  of  CORE  or  EXCORE,  try  again  with  a smaller
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 33


        argument.  A possible remedy might also be to EXCISE and then
        try again.

COULDN'T SSAVE
        LSSAVE failed.

COULDN'T SWAP SUCCESSFULLY
        %SOSSWAP failed.

DOT CONTEXT ERROR
        READ doesn't like dots adjacent to parentheses or other dots.

EXCORE x ON TOP LEVEL REQUIRED
        This error only occurs on a Tops-10 machine.   There  is  not
        enough  BPS for the Fap loader to load code into BPS.  As the
        input is coming from a file, the Fap loader can't use  EXCORE
        to  get  more  BPS,  as  that would clobber the input buffer.
        Execute (EXCORE x) and try again.

FASLOAD BUG
        A bug has been discovered in the Fap loader.

FASLOAD EMPTY FILE
        A file to be loaded by  the  Fap  loader  was  empty  or  the
        fisltable is not a vector.

FASLOAD EXCEEDS BPS
        The Fap loader has run out of BPS.  For a remedy, see message
        under BINARY PROGRAM SPACE EXCEEDED

FASLOAD FORMAT ERR
        Something went wrong during the Fap loading  process.   There
        might be a format error in the Fap file.  This error might be
        preceded by another Lisp  error  message  telling  what  went
        wrong.

FASLOAD GC-PRO ERR This error occurred during the last stage  of  the
        Fap loading process.

FISLTABLE FULL
        The Fap loader needed  a  bigger  fisltable.   On  a  Tops-10
        system,  set FISLSIZE to a larger number and try again.  On a
        Tops-20 or a  Tenex  system,  the  fisltable  should  be  big
        enough,  but it can be enlarged by N words, by the following:
        subtract N from BPORG, add N to FISLSIZE, execute
         (SETQ FISLTABLE (MKVECT (DIFFERENCE (TIMES2 FISLSIZE 2) 1)))
        and set BPORG to its old value.

FLOATING OVERFLOW
        A numeric routine detected a floating point overflow.

GARBAGED OBLIST
        This error is always preceded by some other error.  The error
        message  routine for that error discovered an illegal address
        in the OBLIST or on the property  list  of  some  id  on  the
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 34


        OBLIST.  You are in trouble.

ILLEGAL LAMBDA FORMAT
        APPLY was given a lambda expression whose parameter list  was
        an id other than NIL.

INPUT ERROR
        Bad data was read from the selected input channel.

LISP.ED MISSING
        An ED or GRINDEF command (not  usually  available)  has  been
        issued,  but  the  system  couldn't find the Alvine code file
        LISP.ED.  Either you have defined SYS:  to  be  a  device  on
        which  LISP.ED  isn't found, or there is no such file at your
        installation.

LISP.ED OR LOD MISSING
        Either LISP.ED was not found (see previous message  comment),
        or  a  LOAD  command  has  been issued, but the file LISP.LOD
        wasn't found, in which case the situation is  similar  to  an
        absence of LISP.ED.

LISP.LOD MISSING
        LISP.LOD wasn't found, see previous message comment.

LISP.SYM not found!!  No load
        A LOAD command has been issued, the file LISP.LOD was  found,
        but the symbol table file LISP.SYM wasn't found.

MISSING PRINT NAME IN OBLIST
        INTERN found a member of the OBLIST that has no  print  name.
        You are in trouble.

MORE THAN ONE S-EXPRESSION - COMPRESS
        COMPRESS objects to a list which constitutes  the  characters
        for more than one S-expression.

NO FISLTABLE ROOM
        This can only occur on a Tops-10 machine.  The available  BPS
        is not enough for the Fap loaders fisltable.  You need to get
        more BPS.

NO FREE STG LEFT
        All free storage (Lisp cells) is used.  The garbage collector
        couldn't  find any unused cells.  The unbinding of the stacks
        that occurs  after  this  error  (and  others)  will  normaly
        release some storage.  If that isn't enough, you can get more
        storage by using the CORE function.

NO FULL WORDS LEFT
        All full words are being used for print  names  and  numbers.
        The problem and its solution is similar to FREE STG.
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 35


NO INPUT - RDS
        The argument to RDS was not a channel OPENed for input.

NO I/O CHANNELS LEFT
        OPEN failed to find a free I/O channel.  There is  a  maximum
        of 16 (including the TTY) open I/O channels at the same time.

NOGO
        This usually follows a previous error message  from  the  Fap
        loader.

NO LIST - COMPRESS
        NIL has been given as argument to COMPRESS.

NON-INTEGRAL OPERAND
        DIVIDE, REMAINDER and GCD only accepts integers as arguments.

NO OUTPUT - WRS
        The argument to WRS was not a channel OPENed for output.

NO PRINT NAME
        Some routine couldn't find a print name.  This might occur in
        several ways:

        1.  The argument to INTERN, SETPCHAR or (on Tenex or  Tops-20
            machines)  the  element  after DIR:  in argument to OPEN,
            was not a string or an id.

        2.  The argument to PRINC or SKIPTO  or  an  element  of  the
            argument to COMPRESS was not a string, an id or an INUM.

        3.  The argument to  REMOB,  INTERNP  or  one  of  the  above
            functions, was a string or an id without a print name.


NUMBER NOT DIGIT
        The argument to PRINC or SKIPTO or an element of the argument
        to  COMPRESS  was  an INUM that was either negative or larger
        than 9.

OUTPUT ERROR
        Data was improperly written on the  selected  output  device.
        Possibly a write-locked DECTAPE.

PDL OVERFLOW FROM GC - CAN'T CONTINUE
        There is not enough space on  the  regular  stack  to  finish
        garbage  collecting.  This error forces the Lisp system to go
        to the top level, ignoring any ERRORSETs  that  might  be  in
        effect.  The problem and its solution is similar to the error
        NO FREE STG LEFT.

TOO FEW ARGUMENTS SUPPLIED - APPLY
TOO MANY ARGUMENTS SUPPLIED - APPLY
        APPLY checks that interpreted functions have  correct  number
        of arguments.
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 36


TOO MANY ARGS FOR EXPR
        A function that spreads its argument, an EXPR, was given more
        than 14 arguments.

ZERO DIVISOR
        The second argument to  DIVIDE,  REMAINDER  or  QUOTIENT  was
        zero.



4.2  Error Messages Starting With A Bad Part

The following messages are all of the form:
        ***** x <message> They are ordered alphabetically  after  the
message part.

x CALLED AS EXPR
        This error occurs when a call to "x" from a compiled function
        is  to  be performed and "x" is a FEXPR but the call has been
        compiled as  if  "x"  was  an  EXPR.   This  means  that  the
        arguments to "x" have been evaluated when the call is made.
        At present this error  doesn't  occur.   Instead  the  system
        makes  a  (dubious)  try  to  recover.   It  quotes  all  the
        arguments and makes a list of them and then  calls  "x"  with
        this list as argument.

x CAN'T FIND FILE
        OPEN couldn't find the file specified by "x".   If  you  were
        looking for a file on SYS:, maybe your definition of SYS:  is
        wrong.

x DEVICE NOT AVAILABLE
        The device specified to OPEN in the file name list  "x",  was
        not available.

x DIRECTORY FULL
        The directory is full for the device specified  in  the  file
        name list "x".

x FILE IS WRITE PROTECTED
        OPEN found that the file specified by "x" is write  protected
        and may not be opened for output.

x ILLEGAL ARG TO LINELENGTH
        The argument "x" to LINELENGTH wasn't NIL or a positive INUM.

x ILLEGAL ARG TO PAGELENGTH
        The argument "x" to PAGELENGTH wasn't NIL or  a  non-negative
        INUM.

x ILLEGAL DEVICE
        The device specified in the file name list "x" doesn't  exist
        or  it is of the wrong type.  E.g. it is not possible to open
        the lineprinter for input.
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 37


x ILL MEM REF FROM y
        An illegal memory reference was made.  "x" is the address (in
        octal)  of  the instruction where the error was made.  "y" is
        the function containing that instruction.  "y" is found in  a
        similar  way to the BACKTRACE, so that ?  denotes an internal
        routine and NIL denotes a position outside normal code,  e.g.
        in free storage.
        This error is usually caused by taking CAR or CDR of an atom.

x ILL UUO FROM y
        An illegal instruction was executed at  address  "x"  in  the
        function  "y".   ?   and  NIL have the same meaning as in the
        previous error message.

x IS NOT A CHANNEL NAME
        RDS, WRS or CLOSE was given "x" as argument.  "x" was not NIL
        or an integer between 0 and 15.

x IS NOT AN IDENTIFIER
        "x" should have  been  an  identifier.   The  functions  that
        checks  for  identifiers, are:  PUTD, PUT, FLAG, REMFLAG, SET
        and SETQ.  The elements of the parameter list in a PROG or  a
        lambda expression must also be ids.

x IS NOT A NUMBER
        An arithmetic routine was given an argument "x", that  wasn't
        a number.

x IS NOT A VECTOR
        The second argument to PUTV or GETV must be  a  vector.   "x"
        wasn't.

x MAY NOT BE CHANGED
        "x" is one of the ids NIL and T.  SET and SETQ may not change
        the value of these ids.

(x y) MISMATCHED - PAIR
        The two arguments "x" and "y" to PAIR were not lists  of  the
        same length.

x MODULE NOT FOUND
        FLOAD or FISLM was given the name of a module  that  couldn't
        be found.  Your definition of SYS:  may be wrong.

x NOT A KEYWORD FOR OPEN
        "x" was the second argument to OPEN, but it was  not  one  of
        INPUT, OUTPUT, INBIN or OUTBIN.

x NOT EXPR - APPLY
        "x" was the first argument to APPLY.  It was a  function  but
        not an EXPR.

x PUSHDOWN CAPACITY EXCEEDED
        The "x" pushdown stack has overflowed.   "x"  is  either  REG
        (regular)  or SPEC (special).  The overflow is usually caused
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 38


        by non-termination of recursion.  If this is not the case but
        the  recursion  just  is  very deep, then you can enlarge the
        stacks by using CORE.

x SUBSCRIPT OUT OF RANGE
        The first argument "x" to GETV or PUTV was out of  range  for
        the vector specified by the second argument.

x TOO BIG VECTOR
        MKVECT was given "x" argument, but there is not space in  the
        BPS for a vector of size "x"+1.

x UNBOUND VARIABLE - EVAL
        EVAL tried to evaluate the id "x" and found that  it  had  no
        value.  You probably forgot to QUOTE "x".

x UNDEFINED FUNCTION
        The function "x" is not defined.

x UNDEFINED FUNCTION - APPLY
        The function "x" is not defined.

x UNDEFINED PROG TAG - GO
        A GO in some interpreted function had an undefined label "x".

x UNDEFINED UUO
        The compiled function "y" made a call to "x", but "x" was not
        a  function.  Note that a MACRO is not accepted as a function
        by calls from compiled functions.



4.3  Warnings And Diagnostic Messages

These messages are preceded by a "***".

CAN'T CUT CORE INTO ALLOCATED
        The argument to CORE was to small.  See CORE.

CAN'T EXCISE
        CORE or EXCISE couldn't reallocate or excise  extra  core  as
        EXCORE  had  been  executed earlier with an argument that was
        not NIL.  If you want to excise  or  reallocate  extra  core,
        execute (EXCORE NIL), then try again.

EXCORE x PERFORMED
        This message can only occur on a Tops-10  machine.   The  Fap
        loader  needed more BPS than was available and used EXCORE to
        get more BPS.  This message is not printed if *MSG is NIL.

FREE STG EXHAUSTED
        The free storage (the Lisp cells) was exhausted and caused  a
        garbage  collection.  This message is not printed when *GCGAG
        is NIL.
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 39


FULL WORD SPACE EXHAUSTED
        The full word  space  got  exhausted  and  caused  a  garbage
        collection.  This message is not printed when *GCGAG is NIL.

WOULDN'T REDUCE CORE
        %SOSSWAP couldn't reduce the  core  to  what  it  was.   This
        message  is  only possible on a Tops-10 machine.  %SOSSWAP is
        normally not implemented on Tops-10 machines.

(F)SUBR CONVERTED TO (F)EXPR
        The function types FSUBR and/or SUBR have  been  replaced  by
        FEXPR  and/or  EXPR  by the Fap loader when loading a module.
        This message is not given more  than  once  for  each  module
        loaded.   The  module  was compiled before the function types
        FSUBR and SUBR where deleted from Standard  Lisp.   Recompile
        the  module  and  the  message  will go away.  This automatic
        conversion will removed sometime in the future, an  undefined
        function error will then occur at run time instead.

(F)SUBR CONVERTED TO (F)EXPR IN PUTD
        A call to PUTD used the old function type FSUBR or SUBR  that
        are  now  replaced  by  FEXPR  and  EXPR.   Use FEXPR or EXPR
        instead.  Compiled functions are  now  distinguished  by  the
        predicate  CODEP.  Using (F)SUBR will cause an error sometime
        in the future when this conversion is removed.

x FREE STORAGE, y FULL WORDS AVAILABLE
        There are "x" Lisp cells and "y" full words  available  after
        the  garbage  collection that just occurred.  This message is
        not printed if *GCGAG is NIL.

x REDEFINED
        This message is printed by PUTD or by the Fap loader when the
        function  "x"  has  been redefined.  The message from the Fap
        loader is not printed if *PREDEF is NIL.   The  message  from
        PUTD is not printed if *MSG is NIL.
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 40


5.0  DIFFERENCES BETWEEN LISP 1.6 AND STANDARD LISP

The following is a list of the major differences between LISP 1.6 and
Standard Lisp on the DECsystem 10 and 20:

     1.  There are no LSUBRs and consequently no ARG or SETARG.

     2.  An EXPR may have up to 13 arguments.

     3.  SUBR and FSUBR is replaced  by  EXPR  and  FEXPR.   Compiled
         functions  are  distinguished  from interpreted functions by
         the predicate CODEP.

     4.  There are no arrays, so ARRAY, EXARRAY  and  STORE  are  not
         defined.  Use vectors instead.

     5.  Alvine (ED, GRINDEF) is normally not implemented.

     6.  The Lisp LOADer, *PUTSYM, PUTSYM,  *GETSYM  and  GETSYM  are
         normally not implemented.

     7.  The trace functions TRACE and TRACET  are  replaced  by  the
         somewhat different TR and TRST.

     8.  The I/O functions DSKIN and SYSIN are not defined.

     9.  The functions APPEND, NCONC, DIFFERENCE, QUOTIENT,  GREATERP
         and  LESSP are now EXPRs and take only two arguments.  BOOLE
         takes 3 arguments.

    10.  ASSOC and SASSOC uses EQUAL instead of EQ.  ATSOC is similar
         to ASSOC but uses EQ.

    11.  REMOB is a SUBR with one argument.

    12.  Computed GO labels are not allowed, i.e. the argument to  GO
         must be an id.

    13.  GC, MAKNAM and PRINC are renamed to  RECLAIM,  COMPRESS  and
         PRIN2.

    14.  OUTPUT, INPUT, OUTBIN, INBIN, OUTC and INC are  replaced  by
         OPEN, RDS, WRS and CLOSE.

    15.  The mapping functions (MAP, MAPC,  MAPCAR,  MAPLIST,  MAPCAN
         and MAPCON), take their arguments in the opposite order.

    16.  The functions INITFN, BAKGAG, GCGAG,  NOUUO,  DDTIN,  RAISE,
         FECHO and *RSET are replaced by variables.

    17.  ERRSET is replaced by the SUBR ERRORSET.

    18.  *EXPAND and *EXPAND1 are replaced by EXPAND.
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 41


    19.  The characters "%", "[", "]", "<", ">", "@",  "'",  "/"  and
         "!" have a different meaning.  See RDSLSH.

    20.  The BIGNUM and SCAN packets are permanently resident.

    21.  The basic datatypes are differently implemented.  The format
         for  property  lists  is  changed.  Flags and properties can
         never clash.

    22.  A string is not the same as an uninterned id.

    23.  CHRCT is replaced by POSN.

    24.  *EVAL is renamed to EVAL.  APPLY is an  EXPR  with  2  args.
         %APPLY and %EVAL support the FUNARG feature.

    25.  PROG2 is replaced by PROGN

    26.  The  following  functions  are  not  defined:   CSYM,  LAST,
         DEFPROP, GETL, MAKNUM and UNTYI.

    27.  Several new functions are defined.

    28.  End of line is treated differently.  See ^_ and $EOL$.

STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 42


6.0  REFERENCES


[1] J. B. Marti, A. C. Hearn, M. L. Griss, C.  Griss,  Standard  LISP
                                                       ________  ____
     Report, Utah Computational Physics, UCP-60, January 1978.
     ______


[2] Lynn  H.  Quam,  Whitfield  Diffie,  Stanford  LISP  1.6  Manual,
                                         ________  ____  ___  ______
     Stanford  Artificial  Intelligence  Laboratory,  Operating  Note
     28.7.
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 43


                                Index



$EOF$  . . . . . . . . . . . . . . 28
$EOL$  . . . . . . . . . . . . . . 28

%APPLY . . . . . . . . . . . . . . 17
%EVAL  . . . . . . . . . . . . . . 17

*BAKGAG  . . . . . . . . . . . . . 10, 28
*BOX . . . . . . . . . . . . . . . 18
*COMP  . . . . . . . . . . . . . . 23, 28
*DDTIN . . . . . . . . . . . . . . 28
*ECHO  . . . . . . . . . . . . . . 29
*ERRMSG  . . . . . . . . . . . . . 10, 29
*FUNCTION  . . . . . . . . . . . . 17
*GCGAG . . . . . . . . . . . . . . 29
*MSG . . . . . . . . . . . . . . . 29
*NOPOINT . . . . . . . . . . . . . 29
*NOUUO . . . . . . . . . . . . . . 30
*PREDEF  . . . . . . . . . . . . . 30
*PURIFY  . . . . . . . . . . . . . 30
*RAISE . . . . . . . . . . . . . . 28
*RSET  . . . . . . . . . . . . . . 10, 30

<del>  . . . . . . . . . . . . . . 2

ABS  . . . . . . . . . . . . . . . 8
AND  . . . . . . . . . . . . . . . 8
APPEND . . . . . . . . . . . . . . 8
APPLY  . . . . . . . . . . . . . . 16
ASCII  . . . . . . . . . . . . . . 23
ASSOC  . . . . . . . . . . . . . . 32
ATOM . . . . . . . . . . . . . . . 8
ATSOC  . . . . . . . . . . . . . . 23

BASE . . . . . . . . . . . . . . . 30
BCP (binding context pointer)  . . 17
BIGNUM . . . . . . . . . . . . . . 2
BIGP . . . . . . . . . . . . . . . 8, 22
Binary program space (BPS) . . . . 24
Binding context pointer (BCP)  . . 17
BOOLE  . . . . . . . . . . . . . . 19
BPEND  . . . . . . . . . . . . . . 31
BPORG  . . . . . . . . . . . . . . 31
BPS (binary program space) . . . . 24

CAR  . . . . . . . . . . . . . . . 32
CDR  . . . . . . . . . . . . . . . 32
CLOSE  . . . . . . . . . . . . . . 12
Code-pointer . . . . . . . . . . . 4
CODEP  . . . . . . . . . . . . . . 8, 39-40
COMPILE  . . . . . . . . . . . . . 23
COMPRESS . . . . . . . . . . . . . 22
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 44


COND . . . . . . . . . . . . . . . 23
Cond-form  . . . . . . . . . . . . 23
CONS . . . . . . . . . . . . . . . 8
CONSTANTP  . . . . . . . . . . . . 8
CORE . . . . . . . . . . . . . . . 25

DE . . . . . . . . . . . . . . . . 8
DEFLIST  . . . . . . . . . . . . . 8
DELETE . . . . . . . . . . . . . . 8
DELIMIT  . . . . . . . . . . . . . 21
DEPOSIT  . . . . . . . . . . . . . 20
DF . . . . . . . . . . . . . . . . 8
DIFFERENCE . . . . . . . . . . . . 8
DIGIT  . . . . . . . . . . . . . . 8
DIVIDE . . . . . . . . . . . . . . 8
DM . . . . . . . . . . . . . . . . 8

EJECT  . . . . . . . . . . . . . . 8
EMSG*  . . . . . . . . . . . . . . 10, 28
EQ . . . . . . . . . . . . . . . . 8
EQN  . . . . . . . . . . . . . . . 8
EQUAL  . . . . . . . . . . . . . . 8
ERROR  . . . . . . . . . . . . . . 9
Error handling . . . . . . . . . . 8
ERRORSET . . . . . . . . . . . . . 10
EVAL . . . . . . . . . . . . . . . 15
EXAMINE  . . . . . . . . . . . . . 19
EXCISE . . . . . . . . . . . . . . 26
EXCORE . . . . . . . . . . . . . . 26
EXPAND . . . . . . . . . . . . . . 8
EXPLODE  . . . . . . . . . . . . . 22
EXPLODEC . . . . . . . . . . . . . 22
EXPT . . . . . . . . . . . . . . . 8

Fap loader . . . . . . . . . . . . 18
File name list . . . . . . . . . . 11
FILEP  . . . . . . . . . . . . . . 8, 13
FISLM  . . . . . . . . . . . . . . 18
FIX  . . . . . . . . . . . . . . . 8
FIXNUM . . . . . . . . . . . . . . 2
FIXP . . . . . . . . . . . . . . . 8
FLAG . . . . . . . . . . . . . . . 8
FLAGP  . . . . . . . . . . . . . . 8
FLOAD  . . . . . . . . . . . . . . 18
FLOAT  . . . . . . . . . . . . . . 8
Floating . . . . . . . . . . . . . 4
FLOATP . . . . . . . . . . . . . . 8
FLUID  . . . . . . . . . . . . . . 22
FLUIDP . . . . . . . . . . . . . . 8
FREEZE . . . . . . . . . . . . . . 23
FSUBR  . . . . . . . . . . . . . . 40
Full word space  . . . . . . . . . 24
FUNARG . . . . . . . . . . . . . . 17
FUNCELL  . . . . . . . . . . . . . 6
FUNCTION . . . . . . . . . . . . . 8
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 45


Function cell  . . . . . . . . . . 6

GCD  . . . . . . . . . . . . . . . 22
GCTIME . . . . . . . . . . . . . . 27
GENSYM . . . . . . . . . . . . . . 8
GET  . . . . . . . . . . . . . . . 8
GETD . . . . . . . . . . . . . . . 8
GETV . . . . . . . . . . . . . . . 8
GLOBAL . . . . . . . . . . . . . . 22
GLOBALP  . . . . . . . . . . . . . 8
GO . . . . . . . . . . . . . . . . 32
GREATERP . . . . . . . . . . . . . 8

IBASE  . . . . . . . . . . . . . . 30
Id . . . . . . . . . . . . . . . . 4
Identifier . . . . . . . . . . . . 4
IDP  . . . . . . . . . . . . . . . 8
IGNORE . . . . . . . . . . . . . . 21
INC* . . . . . . . . . . . . . . . 12, 31
INITFN*  . . . . . . . . . . . . . 30
Integer  . . . . . . . . . . . . . 2
INTERN . . . . . . . . . . . . . . 8
INTERNP  . . . . . . . . . . . . . 22
Interpreter  . . . . . . . . . . . 15
INUM . . . . . . . . . . . . . . . 2
INUMP  . . . . . . . . . . . . . . 8, 22

LABEL  . . . . . . . . . . . . . . 17
LENGTH . . . . . . . . . . . . . . 8
LESSP  . . . . . . . . . . . . . . 8
LETTER . . . . . . . . . . . . . . 21
LINELENGTH . . . . . . . . . . . . 14
LIST . . . . . . . . . . . . . . . 8
LITER  . . . . . . . . . . . . . . 8
LPOSN  . . . . . . . . . . . . . . 8
LSH  . . . . . . . . . . . . . . . 19

MAP  . . . . . . . . . . . . . . . 8
MAPC . . . . . . . . . . . . . . . 8
MAPCAN . . . . . . . . . . . . . . 8
MAPCAR . . . . . . . . . . . . . . 8
MAPCON . . . . . . . . . . . . . . 8
MAPLIST  . . . . . . . . . . . . . 8
MAX  . . . . . . . . . . . . . . . 8
MAX2 . . . . . . . . . . . . . . . 8
MEMBER . . . . . . . . . . . . . . 8
MEMQ . . . . . . . . . . . . . . . 8
MIN  . . . . . . . . . . . . . . . 8
MIN2 . . . . . . . . . . . . . . . 8
MINUS  . . . . . . . . . . . . . . 8
MKCODE . . . . . . . . . . . . . . 19
MKVECT . . . . . . . . . . . . . . 8

NCONC  . . . . . . . . . . . . . . 32
NIL  . . . . . . . . . . . . . . . 28
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 46


NOT  . . . . . . . . . . . . . . . 8
NULL . . . . . . . . . . . . . . . 8
NUMBERP  . . . . . . . . . . . . . 8
NUMVAL . . . . . . . . . . . . . . 18

OBLIST . . . . . . . . . . . . . . 31
OPEN . . . . . . . . . . . . . . . 11
OR . . . . . . . . . . . . . . . . 8
OUTC*  . . . . . . . . . . . . . . 12, 31

PAGELENGTH . . . . . . . . . . . . 8, 14
PAIR . . . . . . . . . . . . . . . 8
PAIRP  . . . . . . . . . . . . . . 8
PGLINE . . . . . . . . . . . . . . 13
PLUS . . . . . . . . . . . . . . . 8
PLUS2  . . . . . . . . . . . . . . 8
PNAME  . . . . . . . . . . . . . . 5
POSN . . . . . . . . . . . . . . . 8
PRIN1  . . . . . . . . . . . . . . 8, 15
PRIN2  . . . . . . . . . . . . . . 8, 15
PRINT  . . . . . . . . . . . . . . 8, 15
PROG . . . . . . . . . . . . . . . 8
PROGN  . . . . . . . . . . . . . . 8
Property list  . . . . . . . . . . 5
PUT  . . . . . . . . . . . . . . . 8
PUTD . . . . . . . . . . . . . . . 23
PUTV . . . . . . . . . . . . . . . 8

QUOTE  . . . . . . . . . . . . . . 8
QUOTIENT . . . . . . . . . . . . . 8

RDS  . . . . . . . . . . . . . . . 12
RDSLSH . . . . . . . . . . . . . . 13
READ . . . . . . . . . . . . . . . 14-15, 21
READCH . . . . . . . . . . . . . . 15
RECLAIM  . . . . . . . . . . . . . 27
REMAINDER  . . . . . . . . . . . . 8
REMD . . . . . . . . . . . . . . . 8
REMFLAG  . . . . . . . . . . . . . 8
REMOB  . . . . . . . . . . . . . . 8
REMPROP  . . . . . . . . . . . . . 8
RETURN . . . . . . . . . . . . . . 32
REVERSE  . . . . . . . . . . . . . 8
RPLACA . . . . . . . . . . . . . . 32
RPLACD . . . . . . . . . . . . . . 32

SASSOC . . . . . . . . . . . . . . 8
SCAN . . . . . . . . . . . . . . . 21
SCANINIT . . . . . . . . . . . . . 21
SCANSET  . . . . . . . . . . . . . 20
SCNVAL . . . . . . . . . . . . . . 31
SET  . . . . . . . . . . . . . . . 22
SETPCHAR . . . . . . . . . . . . . 13
SETQ . . . . . . . . . . . . . . . 22
SETSYS . . . . . . . . . . . . . . 13
STANDARD LISP ON DECSYSTEM 10 AND 20                          Page 47


SKIPTO . . . . . . . . . . . . . . 14
SPEAK  . . . . . . . . . . . . . . 27
String . . . . . . . . . . . . . . 4
STRINGP  . . . . . . . . . . . . . 8
SUBLIS . . . . . . . . . . . . . . 8
SUBR . . . . . . . . . . . . . . . 40
SUBST  . . . . . . . . . . . . . . 8
SYS: . . . . . . . . . . . . . . . 13
SYSTEM*  . . . . . . . . . . . . . 31

T  . . . . . . . . . . . . . . . . 28
TERPRI . . . . . . . . . . . . . . 8
TIME . . . . . . . . . . . . . . . 11
TIMES  . . . . . . . . . . . . . . 8
TIMES2 . . . . . . . . . . . . . . 8
TYI  . . . . . . . . . . . . . . . 14-15
TYO  . . . . . . . . . . . . . . . 14

UNBOUND  . . . . . . . . . . . . . 16
UNFLUID  . . . . . . . . . . . . . 22
UNREADCH . . . . . . . . . . . . . 14
UPBV . . . . . . . . . . . . . . . 8

VALUE  . . . . . . . . . . . . . . 6
Value cell . . . . . . . . . . . . 6
Vector . . . . . . . . . . . . . . 4
VECTORP  . . . . . . . . . . . . . 8

WARNING  . . . . . . . . . . . . . 10
WRS  . . . . . . . . . . . . . . . 12

^A . . . . . . . . . . . . . . . . 2
^C . . . . . . . . . . . . . . . . 2
^G . . . . . . . . . . . . . . . . 2, 10
^O . . . . . . . . . . . . . . . . 2
^R . . . . . . . . . . . . . . . . 2
^U . . . . . . . . . . . . . . . . 2
^X . . . . . . . . . . . . . . . . 2
^Z . . . . . . . . . . . . . . . . 15
^_ . . . . . . . . . . . . . . . . 15, 28