Digital PDFs

XX-ED056-25

June 1977

71 pages

Original

6.8MB

Document:	Lick 32K FOCAL Users Guide 197706
Order Number:	XX-ED056-25
Revision:
Pages:	71
Original Filename:	http://bitsavers.org/pdf/dec/pdp8/focal/Lick_32K_FOCAL_Users_Guide_197706.pdf

OCR Text

ALPHABETICAL INDEX OF 32K FUNCTIONS
NAME

PAGE

DESCRIPTION

ADDQ%
ADDR*
ADV
AND
ASK
BAK
BIG
BRK
CADO
CALL
CHAN
CHEK
CHOP*
CLOS
CNT*
COMB*
COMP
CORL*
COTY
CPEN
CRT
CT*
DEC*
DIS
DIVD
DLTQ*
DMUL
DO
DONE*
DTIM*
EDIT
END
EDF
ERAS
EXCR
FILE
FLIP*
GO
HDTS*
HI
HUNT

U10
34
21.3
26.1
14
21.3
26.1

ADD ITEM TO END OF 5 ELEMENT QUEUE
OCTAL ADDRESS OF LAST WORD USED BY X ZAP
ADVANCE IBM TAPE ONE RECORD
AND TWO 24-BIT NUMBERS TOGETHER
READ A FOCAL FLOATING PT. VARIABLE FROM DECTAPE
BACKSPACE IBM TAPE ONE RECORD
CONVERT 12-BIT SIGNED INTEGER TO 24 BITS
BREAK OUT OF A FOCAL DO GROUP OR FOR LOOP
DO A LINE OR GROUP IN ANOTHER PROGRAM
CALL AND OPTIONALLY START ANOTHER PROGRAM
RETURN VALUE OF A SINGLE CHANNEL
RETURN SCANNER/PUSHBUTTON STATUS; UPDATE LEDS
MOVES 120" TELESCOPE
CLOSE OUT DATA-CHAINED DECTAPE BUFFER
START 3-PHOTOMULTIPLIER COUDE DEVICE
COMBINE ADJACENT 12-BIT WORDS IN A BUFFER
DRAW STRAIGHT LINES ON CRT OR CALCOMP
CROSS-CORRELATION FUNCTION ON A BUFFER
PRINT TEXT FROM I.D. BUFFERS ON CRT OR TELETYPE
RAISE/LOWER CALCOMP PEN
PLOT SCANS ON CRT OR CALCOMP
READ AND ZERO SCANNER MEMORY COUNTING TIMERS
CONVERT OCTAL TO DECIMAL
DISPLAY A DOT ON CRT
DIVIDE BUFFER BY ANOTHER BUFFER OR A CONSTANT
RETURN AND DELETE ITEM FROM END OF QUEUE
MULTIPLY BUFFER BY ANOTHER BUFFER OR A CONSTANT
COMPUTED FOCAL DO COMMAND
RETURN STATUS OF QUICK IBM COMMANDS
APPLY DEADTIME CORRECTION TO A BUFFER
SET VALUE OF A SINGLE CHANNEL
RETURN TO CALLING PROGRAM
WRITE AN END OF FILE MARK ON IBM TAPE
SET ALL OR PART OF A BUFFER TO ZERO
EXCHANGE/CIRCULAR SHIFT ROUTINE
SAVE A PROGRAM ON DECTAPE
REVERSE ORDERING OF CHANNELS IN A BUFFER
COMPUTED FOCAL GO COMMAND
RUN HIGH SPEED DATA TAKING SYSTEM
RETURN HI-ORDER 12 BITS OF 24 BIT INTEGER
HUNT FOR DOUBLE END OF FILE ON IBM TAPE

NOTES:
*1
*2

U6
U7

12
17

C3
15

*1
18

33
11
18

34
11

U10
11

U10
29
17

11
21.1
17

20
13

-11

26+1

21.2

* IN NAME FIELD INDICATES COMMAND IS IN AN OVERLAY

SEE DOCUMENTATION FDR 3-PHOTOMULTIPLIER COUDE DEVICE
SEE DOCUMENTATION FOR 32K DATA TAKING SYSTEMS

ALPHABETICAL INDEX OF 32K FUNCTIONS <CONT.>
~1E:3CI=(
!

IBM
.l.H1"ll::.

IEL:O*
IN
IND*
INTP
IPHD*
IPUT
IT1:'.il·\

u :l.~5
•''\•""'I

.-::. .,.·:.

'")'")

.·:•• 1.:••

,.J1'•iP*
,JMS*
!... ED
1...0
1...0G:O*
1...CJOI"\*
1...l=(EC

26. :I.

MCEN*
MCCH

MEMC
MEME
MEMR
MEMW
MEMX*
MEMY>l<
MGET

U :L:L

32
..., ...
)

. :> . ...

MO~JE

MP><*
i"1l~:C~*

U:l.2

M~)1~1V

:I. '7

Nr~ME

:1.2

OCT*
OUT

::~4

P1:'.\I<~*

U:l.2

F' i'.1U'.~)
F'E1~1l·\

:l 7

PIXL*
i::"CJLY*
i::·osN*
PUT
PUTN

U:l.4

NOTES:
*1
*2

I PT I DN

:1.4
17

RE"fl.JRN IBM TAPE STATUS
ERASE 4 FEET OF IBM TAPE
RETURNS FIELD DF LAST WORD USED BY X ZAP
ADD 2 BUFFERS TOGETHER
FIRST MOMENT PEAK FINDER
RETURN AND SET MASK OF P/S MONITOR
DISPLAY VARIABLE DENSITY PIXELS ON CRT
WRifE A :1.2 OR 24 BIT WORD TO I.D. BUFFER
RETURN 12 OR 24 BIT WORD FROM I.D. BUFFER
SIMULATES A PDP-8 JMP INSTRUCTION
SIMULATES A PDP-8 JMS INSTRUCTION
DISPLAY NUMBERS ON LEDS
RETURN LO-ORDER :1.2 BITS OF A 24-BIT INTEGER
TAKE :l.OOOO*LOG BASE :1.0 OF A BUFFER
PLOT INTENSITY MAP ON CRT <FOR MAPPING TUBE>
RETURN RECORD LENGTH/LOG I OF LAST IBM RECORD
SET ~3WEEP CENTEI=<
START SCANNER COUNTING WITH EXTERNAL CLOCK PULSE
START SCANNER COUNTING OR RETURN REMAINING TIME
ERASE ALL OF SCANNER MEMORY
READ SCANNER MEMORY WITHOUT ERASING IT
WRITE SCANNER MEMORY
LOAD/READ X SWEEP
LOAD/READ Y SWEEP
READ A 2048-CHANNEL SCAN AND I.D. FROM DECTAPE
COPY FROM BUFFER TO BUFFER
TRANSMIT/RECEIVE FROM/TO SERIAL MULTIPLEXER
READ SCANNER MEMORY, ERASING WHAT IS READ
WRITE A 2048 CHANNEL SCAN AND I.D. TO DECTAPE
READ AND ACTIVATE FOCAL COMMANDS FROM OVERLAY
CONVERT FROM DECIMAL TO OCTAL
SUBTRACT A BUFFER FROM ANOTHER
PACK LO PART OF N CHANNELS IN N WORDS
~:)T(.~h:T/~3TDP sc·ANNEI~: Fl=(OM COUNTING; NO TIME CHr~NGE
FIND PEAK IN A DUFFER
SET SCREEN PARAMETERS FOR X IPHO
GENERATE POLYNOMIAL IN A BUFFER
RETURN POSITION COORDINATES OF :1.20" TELESCOPE
WRITE A :1.2-BIT WORD TD DECTAPE
LOAD A BUFFER WITH LINEAR DATA

* IN NAME FIELD INDICATES COMMAND IS IN AN OVERLAY

SEE DOCUMENTATION FDR 3-PHOTOMULTIPLIER COUDE DEVICE
SEE DOCUMENTATION FOR 32K DATA TAKING SYSTEMS

ALPHABETICAL INDEX OF 32K FUNCTIONS <CONT.>
Ii::· TI ON

N1~!'i[

Pt:~GE

:OE~:;CF(

1:<1~'.:lT::l<

MOVE :1.20" TELESCOPE WHILE SCANNER IS COUNTING
READ 3-PHOTOMULTIPl...IER COUDE DEVICE
QUICK IBM READ COMMAND
READ A RECORD FROM IBM TAPE
REFORMAT BUFFER FROM/TO 32K/BK FORMAT
RETURN POSITION OF X-Y STAGE
1;;:EWIND IBM TAPE
SET COEFFICIENTS FOR X POLY
SCRUNCH/EXPAND SCANSv CONSERVING COUNTS
FIND AL.I... PEAKS IN A BUFFER
SET FUNCTION BUFFER IN 3-PMT COUDE DEVICE
SHIFT BUFFER BY INTEGRAL/FRACTIONAL CHANNELS
COMPANION FUNCTION TO FIND
TRANSMIT/RECEIVE TO/FROM SPECTROGRAPH CONTROL
MOIJE X·-Y ST1~GE
SWITCH OUTPUT BETWEEN TELETYPE AND CRT
STEP 3-PHOTOMULTIPLIER COUDE DEVICE
SAVE A FOCAL FLOATING PT. VARIABLE ON DECTAPE
READ SWITCHES; LIGHT LAMPS; ERASE CRT
READ A 12-BIT WORD FROM DECTAPE
READ TIME/DATE FROM TIME STANDARD
CORRECTION FOR ATMOSPHERIC EXTINCTION
TOTAL OF CHANNEL CONTENTS IN A BUFFER
TRANSLATE DATA THRU ARBITRARILY DEFINED FUNCTION
RETURN POSITION ANGLE OF TUB
STORE CHARACTERS FROM TELETYPE INTO I.D. BUFFER
UNPACK N 12-BIT WORDS INTO N CHANNELS
ERASE UNWANfED FOCAL VARIABLES
LIST OVERLAYS ON DECTAPE
WRITE A RECORD TO IBM TAPE
QUICK IBM WRITE COMMAND
IN'.:)PECT/MODIFY r~L.L ~321"\ OF MEMOl:;:y
ZERO CALCOMP PLOTTER ORIGIN

F<CNT;:<
i:(DQ*
1:(E1°~D

F<EFM
f<TCL*
F(l.llND
SAV»:
~3Cl:;;N*

'.3EEI"\*
'.3ET

:~<:I.

UB
2:1.
:1.9

*2
2:1. ....·.

'")

:~·?

3:1.
~5 :I.

*:I.

~>HOV

:1.6

~:;I G*

~;)PEC>K

;{<2

'.:lT1'iG*

>K::~

STt1T

1 :I.
*:I.
:1.4
1 :t
:1.4

~:>TEF·

STOI:;;
!:;WIT
TAI"\
TIME*
TINC>K
TOTI...
TF((.~N>K

TUB*
TYCO
UNPI"\
lJAF<
l,JHr::iT
~,Jf(

32
29
:I. "7
U:l.!'5
*::.~

... ... )

...·......_

U:l.3
:I. :I.
:l. :I.

WF<C~*

U'?

z,:1Px<

34
:I. fl

ZCOM
NOTi:::'.3:
>:<:I.

>K2

.....

2:1. +,fl,. . .

IN NAME FIELD INDICATES COMMAND IS IN AN OVERLAY
SEE DOCUMENTATION FOR 3-PHOTOMULTIPLIER COUDE DEVICE
SEE DOCUMENTATION FOR 32K nATA TAKING SYSTEMS

Ackno;.;rledgments

In addition to Jack Baldwin, who did the bulk of the original system design, I would like to thank the following people:
Sandy Faber and Heywood Sobel, who, as the first users of the
system, helped to locate a number of weak points in the system that
have now (hopefully)been eliminated.

Dave Burstein, Steve Grandi, and

Ed Kemper, for their advice on deadtime corrections, peak finding, and
IBKtapes, respectively.

Ted Cantrall, for his suggestions on having

32K Focal adapt itself to various size machines.

Alan Koski, who suggested

that 32K Focal should be able to handle scans longer than 2048 channels.
And finally, Lloyd Robinson, whose suggestions are too numerous to elaborate,
and who always managed to find something else ne;; to add to the system.

ABSTRACT
This report is designed as a User's Glide to the Lick 32K Focal
System.

It assumes some familiarity with Digital Equipment Corporation's

Focal 1969 language.

Although this report makes numerous comparisons

with the Lick 8K Focal System (which is described in previous Lick Technical
Reports), an understanding of Lick 8K Focal is not a prerequisite.
The first part of this report describes the similarities and differences between the Lick 8K Focal System and the Lick 32K Focal System.
It also describes some of the new features and concepts one needs to
understand to efficiently use the system.

The second part describes the

operation of specific groups of 32K Focal functions, and serves as an e:l!!tended summary of all the commartds that are a resident part of 32K Focal.
The third part describes additional commands that are available as overlays.

Appendix A gives a seriesof indices and cross-references of Focal

function names and related codes.

Appendix B gives a complete list of

error codes and their associated meanings.
It is probable that updated versions (or additional pages for this
version) will appear from time to time.

** Appendix A was deleted when this report was amended in June 1977.

INDEX

II.

Introduction
a) Background

b) Mtjor Differences Between Lick 8K and 32K Focal

c) Similarities Between 8K and 32K Focal

d) The Five Core-Data Buffers

32K Focal Resident Functions
a) Equivalent functions

b) Program/Name functions

c) Buffer independent Dectape I/O functions

d) Single buffer functions

e) IBM tape functions

I.D. functions

21. 9

III.

IV.

g) Scanner memory control functions

h) Double buffer instructions

i) Notes and miscellaneous instructions

32K Focal Overlay Functions
a) Description

b) Polynomial Overlay

c) Deadtime, extinction, and logarithm overlay

d) Peak finding and linearization overlay

e) Sweep, time, position overlay

f) Cross correlation overlay

g) Debugging utility overlay

h) Intensity map overlay

32K FOCAL update - June 20, 1977

APPENDICES

B-1

Error Diagnostics of Focal, 19 69

B-2

32K Focal Resident Function Error Diagnostics

B-3

32K Focal Overlay Function Error Diagnostics

32K FOCAL USER'S WIDE

Introduction

Background
To assist in data acquisition and data reduction, three PDP SI computers

were installed at Lick Chservatory in the early 1970's.

Since these computers

were to be programmed by astronomers and researchers and not by programming
specialists, a programming language was needed that was:
1.

Easy to learn

Simple to use

Able to perform specific operations on large arrays of nunbers at
high speed

Ab le to interact with a wide variety of specialized electronic hard-

ware.
Digital Equipment Corporation, the manufacturer of the PDP-SI, supplied
an

interactive programming language called Focal 1969, which met the first

tlYo criteria.

Lick then made extensive modifications to Focal so that it

satisfied the latter two.

The end product of those modifications, Lick SK

Focal, is extensively documented in previous Lick Technical Reports.
Although Lick SK Focal has evolved considerably since its inception,
its gro;vth in certain directions has been constrained by a lack of core
memory.

Despite such useful features as the extended disk buffer, many

operations involved in processing scanner data remain cunbersome and
awkw"ard to program.

Further, as the volume of data to process has grown, the

PDP SI' s have become increasingly backlogged.

But without additional core

memory, little could be done to improve Focal.
When less expensive add-on core memories for the PDP-SI became available
in the mid-1970's, the idea of developing a more efficient Focal became
feasible.

The decision was made to upgrade the PDP-SI's to their maximum

memory capacity of 32K of core, and to expand Focal to utilize this additional memory.
tinued

Work on 32K Focal began in 1975 ·by

Jack Baldvin, and con-

through M3.y 19 76 at which point it was taken over by Bob Kib rick.

This report describes the result.

Mljor differences between Lick 8K Focal and Lick 32K Focal

Where SK Focal had 2 core data buffers of 512 channels each, 32K

Focal has 5 core data buffers of 204S channels each.

As a result, it is much

simpler to manipulate 2048 channel scans in 32K focal.
2.

In the 8K Focal System only a few Focal functions could be in core

memory at any one time.

Because of this space limitaton, the Focal functions

were stored on the disk.

If a Focal function was not already in core at

the time it was needed, a copy of it was read from disk to core.

Read time~ 100 msec.

In the 32K Focal system, all of the Focal functions (excepting those
brought in with the X NAM: command) are core resident for faster access.
3.

The text area for Focal programs has been expanded so one can write

larger programs and thus not have to do as much chaining between programs.
4.

The DF 32 disk is no longer

used.

Data can nOW" be transferred

directly from buffers to IBM tape without having to be written onto the disk.
5.

The Lick Focal functions are nOW" faster and simpler to use.

X NAM: overlays written for the SK system cannot be used on the

32K system, and vice versa.
7.

F<X:al programs written for one system generally will not work the

same on the other system.
S.

'J.Wo X NAM: overlays can be in the core memory at the same time.

The SK Focal and 32K Focal systems record scans on Dectape in

different formats, although they both use the same amount of space for each
scan.

10.

A total of 36 scans of 204S channels each could be stored on a

Dectape under the 8K system; 46 scans can be stored on a Dectape using the

32K system.

As a result, the scans are numbered differently inthe 2 systems.

8K scan O starts in block 500 octal (320 decimal) while 32K scan 0 starts in
block O.

SK scan #N is located on the same place on the Dectape as 32K

scan #N+lO.
11. In the SK Focal System, if one erased the CRT and then immediately
tried to write on the screen, the resulting image would be too faint.

a result, one had to insert Focal delay loops to wait 0.5 seconds between
CRT erase and CRT write commands.

In 32K Focal, all commands which write

on the CRT check a hardware flag and wait until the CRT has finished erasing
before beginning to write on the screen.

12.

In the 8K Focal System, if a Lick Focal function was called with

invalid arguments or asked to operate on invalid data, the function would
often proceed with its calculations and return a result as if nothing were
wrong.

This would make program debugging

difficult.

In the 32K Focal

system, wherever possible, if a Lick Focal function detects invalid arguments
or data, it will print
c)

a specific error message and stop execution.

Similarities Between Lick 8K Focal and Lick 32K Focal
1.

The layout of the program Dectape is very much the same for both

systems.

X NA:IB Overlay 1 begins on block 134 (octal); Program 0 begins in

block 160 (octal).

Thus, the Focal tape copiers can still be used on 32K

Focal tapes.
2.

The 32K Focal system and 8K Focal systems record scans on tape in

different formats, although they both use the same amount of space for each
scan.

However, the 32K system can read scans written in either format and

will automatically perform

any needed conversions.

also write scans in either format.

The 32K system can

At present, (Jan/77) the 8K system can

only read and write scans in 8K format.
3.
d)

The bootstrap procedure for 32K Focal is the same as that for 8K Focal.

The 5 Core Data Buffers
The object of 32K Focal is to make processing scans of 2048 channels

both faster and easier.
a

Five core data buffers are available, each with

capacity of 2048 channels.

'IWo different modes of addressing allow one

to treat these buffers as 5 separate 2048-channel buffers, or to treat
groups of these buffers as if they were a single, longer buffer.

For

example, 2 of the buffers can be treated as if they were a single 4096channel buffer; or 5 of the buffers could be treated as if they were a
single 10,240 channel buffer.
CHANNEL
0

BUFFER 0

2 0 4 7 2 048

40 9 5 4 0 96

S 14 3 6144

8 I 9 I 8192

I 0 23 9

2047 2048

4095 4096

6143 6144

2047 2048

4095 4096

6143

2047 2048

4095

. _I_____.....I____-"'-1-----1------~
........_....................~
. . .0

BUFFER
B'UFFER 2

BUFFER 3

·o BUFFER 4'

2047·

The above diagram illustrates hew the same area of memory can be addressed
in a variety of different ways.

Specifying channels 0-2047 of buffer 4

(hashed in area) is equivalent to specifying channels 6144-8191 of buffer 1,
or channels 8192-10239 of buffer 0.

The different addressing modes are best

illustrated by example.
FTOTL is a Focal function which sums up the counts in a specified
number of channels.

The form of the FTOTL command is:
Buffer nurrb er
~Starting
~~Nunber

channel nunber

of channels to process

Set D = FTOTL (A,C,N)
To get the sum of the counts in channels 1-20 of buffer 0, one could code
Set D = FTOTL (0,1, 20).

To get the sum of channels 2051-2055 of buffer 0

one could code Set D = FTOTL (1,3,5), since channel 2051 of buffer 0 is
equivalent to channel 3 of buffer 1.
It is often the case that one wishes to treat the buffer area as 5
distinct buffers of 2048 channels each, and to have a command operate on
all 2048 channels of each such buffer.

To simplify coding in these cases,

if the starting channel number (C) and the number of channels to process (N)
are both left to default to zero, then the first 2048 channels of a buffer
are processed; that is
Set D = FTOTL (A) is equivalent to Set D = FTOTL (A, 0, 2048).

the case wtere N is zero but C is not, the processing begins in the channel
specified by C and continues through channel 2047 of that buffer; that is
Set D = FTOTL (A,C) is equivalent to Set D = FTOTL (A,C, 2048-C)
(to apply this rule in cases where C > 2047, adjust A to bring C

into the range O <= C <= 2047, i.e., Set D = FTOTL (0,2051) is

equivalent to Set D = FTOTL(l,3), which is equivalent to Set D =
FTOTL (1,3,2045).
On the other hand, one may often want to operate on scans longer than
2048 channels, and that is the reason for the other mode of addressing.

To get

the sum of channels 0-4095 of buffer 0, one would code Set D = FTOTL (0,0,4096).
To sum channels 4096-10239 of buffer 0 one would code Set D = FTOTL (O, 4096,
6144) or equivalently Set D = FTOTL (2,0,6144).

In this last case, one is

effectively treating the buffer area as if it were only 2 long buffers -- buffer
0 which has 4096 channels and buffer 2 which has 6144 channels.

Note:

operations

need not begin or end on channel nwnbers which are even multiples of 2048.

sum up channels 2001-10000 of buffer 0, one would code Set D = FTOTL (O, 2001, 8000).
A few words of caution are in order on this point.

The dual mode of ad-

dressing allows one to concisely refer to scans of exactly 2048 channels while
at the same time providing the flexibility to manipulate in a single command
scans in excess of 2048 channels.

However, the ability to refer to the same

area of memory in more than one way also makes it easier to get confused and to
try to use the same area of storage at the same time for mutually exclusive
purposes.
For example, if one wants to maintain a long scan of 6144 channels in channels
0-6143 of buffer 0, he must remember that in essence he is also occupying channels
0-2047 of buffers 1 and 2.

Buffers 3 and 4 are still safe to use in this case

as they do not overlap with the area in which the long scan is being stored.
If one is only working with scans of 2048 channels and as long as there is
no need to specify a non-zero number of channels (N), then different buffer nwnbers
will never refer to overlapping areas of memory.

That is, Set D = FTOTL (A,C)

and Set D = FTOTL (B,C) will always reference mutually exclusive areas of memory
(provided A~ B, of course!).
It is also possible to specify combinations of buffer nwnbers (A), starting
channel nwnbers (C), and numbers of channels to process (N), which are plainly
illegal, and some which perhaps should be legal but are not.

Here are some

examples of INVALID commands:
Set D = FTOTL (5)

- Buff er 4 is highest legal buffer number
(on a machine with 32K memory - see next
section).

Set D = FTOTL (3,5100)

The highest legal channel in buffer 3 is 4095.

Set D = FTOTL(3;6144,10240)- Nice try but it won't work.

Although channel

6144 buffer 0 is equivalent to channel 0
buffer 3, channel -6144 buffer 3 is not equivalent to channel 0 buffer 0.

Negative chan-

nel numbers are completely taboo!
Set D = FTOTL (0,2047,-2048)-Nice try again~ no you can't start at the
righthand side and count backwards to the left.
Negative numbers of channels to process are also taboc

If you use any of these illegal varieties of buffer specifications, Focal will
print an error message and refuse to execute the command.

(Error messages

are listed in Appendix B.)
It is also possible to come up with specifications of A,C, and N which,
although not really proper, will be executed without giving an error message.
As an example
Set D = FTOTL (3,0, 8192) is not really correct in that there are only
4096 channels available if one starts at channel 0 of buffer 3.

No error mes-

sage is given and the command is treated as if it had been Set D = FTOTL (3,0, 4096).
As a final note on the data buffers, it should be noted that although Lick
32K Focal is designed to operate on a PDP-8 with 32K of core memory, it can still
be used on PDP-8's with less than 32K of core.
buffers are available.

If less core is available, fewer

The following table relates the number of available buf-

fers to the amount of core memory available.
Amount of Memory

Available Buffer Numbers

32K

0, l, 2, 3, 4

28K

0, 1, 2, 3

24K
20K

0, 1' 2
0, 1

16K

12K

No buffers available
Extended instructions not usuable

No buffers available
Extended instructions not available
Only 1 X NAME buffer available

Nothing available.
4K
When Lick 32K Focal is bootstrapped, it automatically determines the amount
of core memory available on the machine, and adjusts itself accordingly.
to

Attempts

use buffers that are not available will be diagnosed by Focal as errors.

Thus, the command:
Set D = FTOTL (2)
will work correctly on machines with 24K or more of core, but will cause an error
message to be printed if used on a machine with less than 24K of core.

11
II.

32K FOCAL Functions

a) Eouivalent Functions
The following Lick FOCAL functions work the same in both 8K and 32K FOCAL
and are always available:
X CPEN(P,T)------------ P = 0: Pen Up. P = 1: Pen Down. Pause for
a time about lO*T msec. Pen motion needs
about 100 msec, which can be used for
computation, or by the pause.
X DIS(X,Y)-----------Store a dot on the CRT at location (X,Y).
Full scale is 1023.
X GO(S,L)~-------------~Like ordinary GO, DO but with computed arguments.
X DO(S,L)
(Subroutines, line L.)
X END(O) ~-------------Return to calling program; next line
,..---------Quantity to be displayed (2 23 ~N '2: -2 23 )
,-----------LED number for lowest precision digit (1-8)
,-----------Number of digits
,..-------Non-zero to display blanks
Non-zero to suppress overflow warning
,---------Non-zero to suppress negative number check
X LED(X,Ll,NL,B,O,M)

Display numbers on LED digits on switch panel

, - - - - - - - - X origin #- 0
origin#- 0
/;
,------------------Letter size

,.--------------y

X STAT(X,Y,S)

Direct all future printing to the CRT. Redirect
printing to teletype if X = -1 or for CTRL·C, or
for any error diagnostic.

X SWIT(·l) - - - - - - - - - Erase CRT (Can be followed immediately by CRT write).
X SWIT(O,L)--------------------Load lamps L. Lamps are coded 1,2,4,8,16,32.
SET D = FSWIT(N,S, O, O,M) -----Read switch N, S to D. Set M=4095 to read all of
group Nat once (Use 4094 to read all of group 3).
M=l29 would read switch 1 & 8, weighted. M = 0 to
read only switch N,S.
SET D = FSWIT(3, 11,X, Y) -----Display joystick marker on CRT at (X, Y). When
switch 3,11 is pushed, return 1024*Xl + Yl
where (Xl,Yl) is final marker location on CRT.
SET D = FVAR(O) - - - - - - - - S e t D to end point of FOCAL variable list.
X VAR(D)---------------------Erase all variables defined beyond point D in list.
X WHAT(N,M)--------------Type the names of M user generated overlay programs
as found on program DECtape, starting at overlay :ffN.

b) Program/NAME Functions
The following Lick Focal Commands are also always available, but are slightly
different than their SK counterparts
X NAME(N,F)

- Reads into name buff er #F and makes
available for use the commands from
user generated machine language overlay
#N. There are 2 name buffers (numbered
1 and 2) on machines with 12K or more
of core; only name buffer 1 is available on machines with less than 12K of
core.
Note: A) F ~ 2 is treated as if it
were F = 1
B) X NAME(O) is treated the same
as X NAME(l), except that X NAME (0)
will always cause a fresh copy of overlay #1 to be read in from tape. The
commands in NAME overlay 1 are read
into name buffer 1 at bootstrap time.

X CALL(N,S)
(For greater versatility in
calling, see the X CADO
command, pg. U7)

- Call Program N from Dec tape. If S.> O,
starts program N at subroutine s.
(Code S*l28 + L to start program N at
line L of subroutine S). Use X END(O)
to return to the line following the
X CALL command in the calling program.

Important Notes:
A) It is illegal to use either X CALL or X END from inside a DO group
or FOR loop.
B) If FOCAL's text area contains new or modified text that has not yet
been filed (See X FILE command below), then execution of X CALL or
X END will cause FOCAL to type "REALLY?" on the teletype, and then
to wait for a reply. Any reply other than "Y" will cause the X CALL
or X END command to be supressed, and the contents of the text area
to be preserved. A reply of ''Y" will cause the command to proceed
and the contents of the text area to be lost.
C) If X CALL is used to address a program area on the Dectape that does
not contain a FOCAL program, the following action is taken:

1) If the program area addressed by X CALL
is not completely empty, the error message
BAD PROGRAM nnnn
is printed, where nnnn is an octal number
indicating the number of non-empty (i.e.,
non-zero) locations in the program area.
Focal's text area is erased, and is marked
as not containing any program.
2) If the program area addressed by X CALL

is completely empty, no error message is
printed.

Focal's text area is erased,

and marked as containing the requested
program.

FILE(N)

File program N on Dectape.

If program N

was not first called in using X CALL (N),
the message

OK?
will be printed. If the previous program
(or perhaps old data on the tape) is to
be overwritten, type "Y." To avoid writing,
type "N" or CTRL-C. The program is still
intact in core after being filed (or not
filed) on tape. X FILE also prints last
address used for text;highest address=l2567.
A note on adding new programs:
Before one starts typing in a new program, one should decide which program number to try to file it under.
program number is in use.
in use, type X CALL(N).

Next, one should check to see if that

To check if the area on the tape for program N is
If this area on the tape contains data or an over-

lay, one will get the message
BAD PROGRAM nnnn
If this happens, another program number should be used (unless one doesn't
care about losing the data or overlay).

If one doesn't get the message

BAD PROGRAM nnnn
then the area of tape referenced by X CALL(N) either contains a valid Focal
program or is empty.

Type W to list the program.

Continue hunting for an

available program number until you find an empty portion of tape or until
you find a program that you no longer need.

c) Buffer Independent DECtape I/o Functions
X PUT (W,B,U,I)

- Store integer I in word W of Dectape
block B on Dectape unit U.

SET I = FTAK (W,B,U)

- Return single precision value of word W
in Dectape block B from Dectape unit U
(Note integers have values of 0 <=I<=4095)

X STOR(W,B,U;V)

- Store variable V starting at word W of
Dectape block B on Dectape unit U.
(Note Semicolon!)

SET V = FASK(W,B,U)

- Return 10-digit floating point format
variable starting at word W in Dectape
block B from Dectape unit U
(Note - each floating point variable occupies 4 words)

Note; on the use of W,B, and U in PUT, TAK, ASK, STOR
1.

There are 129 words in each block numbered 0-128

There are 1474 blocks on a Dectape, numbered 0-1473

If one counts total words on a Dectape, starting at word O block 0, there
are 190146 words, numbered 0-190145.

The location of a given word on the

Dectape can be specified in a number of different ways.

For example, word

128 block 1473 is equivalent to word 190145 block O; word 0 block 1 is equivalent to word 129 block 0.
4.

If B=W=O, the previously used Dectape location will be incremented (by 1 for PUT
and FTAK; by 4 for STOR and FASK) and taken as the Dectape location to be used
for the current command.

However, if the unit specified (U)is not the same as the

immediately previously used unit, an error message will be given.
5.

One block of the Dectape is usually kept in core.

If a negative unit number is

specified,(use -8 for -0 ), and the desired block is already in core, the Dectape
will not be re-read or rewritten.

A block in core will only be saved in

this case when a different block needs to be brought in. This process is called
data chaining and saves time since the Dectape doesn't have to be read and written
as often.

If a positive unit number is used, the specified block is always re-read
from the Dectape, regardless of whether it is already in core. In the case of
PUT and STOR using positive unit numbers, the desired block is read into core,
modified, and then immediately written back out to tape.
Although data chaining is faster, it is more dangerous and should be used with
caution. Since a block that has been brought into core and modified is not
written back out to tape until a different block is read in, one could lose
information and also unintentionally write on the wrong tape if one switches
tapes while data chaining is in progress. To avoid this problem, and to also
provide other useful information, a companion command to PUT/TAK/ASK/STOR has
been implemented, called CLOS.

X CLOS(O)

X CLOS(-1)

SET D = FCLOS(l)

- Closes the Dectape buffer. That is, if the
core buff er contains a chained tape block that
has been modified and needs to be written to
tape, this will cause it to be written. IF
DATA CHAINING HAS BEEN USED, 111IS COMMAND SHOULD
BE GIVEN BEFORE SWITCHING TAPES OR UNITS.
- Purges the Dectape buffer. That is, if the core
buffer contains a chained tape block that needs
to be written to tape, X CLOS(-1) will turn off
the data chaining flag and make it appear as if
the block had been written to tape, without
actually writing it. The contents of the Dectape
buffer remains the same. X CLOS(-1) should
not normally need to be used unless you get into
trouble.
- D is set to the absolute word address of the
next available word on the Dectape. X PUT(l27,
1473) followed by SET D = FCLOS(l) would cause
D to be set to 190145.

It should be noted that certain other commands use the Dectape buffer for
scratch space - in particular X WHAT uses the buffer. These commands effectively execute an X CLOS(O) before they begin execution.

16
A Note on Write Protection
Lick 32K Focal has a write protection feature that prevents conunands
like X PUT and X STOR from writing over. the Focal system, the user generated
machine language overlays, or the user's Focal programs up to program #65.
Only write operations to unit 0 are protected -- writes to any other unit can
reference any block.
Different Focal tapes may have different numbers of blocks protected.
Currently, Jan./77,(and this is likely to change) write attempts below block
640 decimal (block 1200octal or start of program -66 or start of scan 20) are
prevented. Please note that the physical reel of tape itself is not protected -- it is only protected when it is mounted on unit O; the same physical reel of tape is unprotected if it is mounted on a unit other than 0.

d) Single Buffer Functions
The following instructions are available only on machines with 16K or
more of core. Although these commands operate on only one buffer, this buffer
can be up to 10,240 channels long. Remember-

for functions of the form

X ~(A, ••• ,C,N), unless otherwise noted, the following compact forms apply:
X ~(A) is equivalent to X ~(A, ••• ,0,2048)
X ~(A, ••• ,c) is equivalent to X ~(A, .... ,c,2048-C)

(See section I-d., "The
Five Core Data Buffers",
on page 7)

Buffer
,..----Positive or negative shift
(Fractional shift in units of l/4096th
of a channel)
- - - - - - Starting channel
Number of channels

X SHOV (A, 4096*0,C,N)

A(x) = A(x+D), where A(x) means channel x
of buffer A.. Shifts scan by both integral
and fractional amounts. If N=O, channels
shifted past channel 0 or 2047 are lost.
If N~O, channels can be shifted past
buffer boundaries. For example, X SHOV(0,8192,20)
will shift channels 20 to 2045 of buffer 0 into
channels 22 to 2047; the original contents of
channels 2046 & 2047 are lost. X SHOV(0,8192 1 20,
2048) will shift channels 20 to 2067 into channels
22 to 2069.

---Buffer
Starting channel
Number of channels
Returns contents of peak channel if = 0
Returns peak channel # if > 0
Returns # of monotonically increasing channels
from C if< O
If non-zero, ignore channels whose content
/
is 1ess than DEF.
S D = FPEAK(A,C,N,MODE,DEF)
---Buffer
/"""~Starting

channel
/!';-----Number of channels
S D = FTOTL(A,C,N)
Total of counts in N channels

,..------ Buffer
~----------

~
X ERAS(A,C,N)

Starting channel
Number of channels
A(X) = O; for X = C to C+N-1

/Channel
//Buffer
S D = FCHAN(C,A)
D = A(C) returns value of specified channel

to store
set value of specified channel
,..---- Buffer
- - - Increment
Initial value
/ /
Starting channel
//Number of channels
X PUTN(A,INC,XO,C,N)
Load N successive channels with lineardata:
X'O, XO+INC, X0+2*INC, .... XO+(N-l)*INC

S D = FZCOM(R)

Resets plotting or1g1n to R returns previous
plotting origin. Used with'X CRT to plot scans.
Plotting origin is set to 4095 bootstrap.
Buffer
Double precision vertical scale
Double precision vertical offset (+ or -)
of CRT sweeps if >O (default 3 )
Calcomp without dots if -1
Calcomp with dots if ~2
Points/chan if >o
Chans/point if <O (default 2)
Starting channel
f /Number of channels

;';--starting X value on CRT
X CRT (A,SC,OR,SW,P,C,N,XO)

Display.scan on CRT or Plot on calcomp
Note: if P > 0 and P <= 4, then each
channel value will be displayed using P
adjacent dots on CRT. For P> 4, only 4
adjacent dots will be displayed. (Note, to
display P adjacent dots when P > 4, code
P+4096)

X distance
Y distance
=0 for line Y, then line X
~O for diagonal line
=O for Calcomp plotter
~o for CRT
X origin on CRT
Y origin on CRT
X COMP(X,Y,D,SW,XO,YO)

Draw a line Y units in Y,
then X units in X. Move diagonally
i f D ~ O.

IMPORTANT NOTE ON CALCOMP PLOTTING
In order to speed up data taking operations, the X CRT and X COMP commands
have been modified so that they can be interrupted by the scanner memory (see
App end ix C) •

Buff er

(BUFFER SPANNING NOT ALLOWED)

Scan number (0-45)
Dectape unit (0-7)

~Format selection (8K Fo:nnat if~ 0)
/ / / E r r o r control (returns -1 for error i f ~ O).
Set D= FMSAV(A, S, U, M, E) Saves 2048 channel buffer A and its 31
word ID area as scan S on Dectape unit U.*
Translates to 8K scan format if M I- 0;
otherwise writes without translation. If
E f 0 returns control to focal with a
value of -1 if tape error occurs; otherwise types TAPE? TAPE? ETC. (Note: unit
0 is write protected against MSAV - see
PUT and STOR) also note: if MI- 0 the
buffer is left in BK scan format after
the write has completed. To translate
buffer back to 32K format, use X REFM command.
,.----Buffer

(BUFFER SPANNING NOT ALLOWED)

- - - - Scan number

~~~:~at

~~;---Error control

Set D = FMGET (A, S, U, M , E)

Reads scans from dectape unit U into
2048-channel buffer A and its 31 word
I.D. area~· If M= 0, performs format
conversion automatically if needed. If
Ml- 0, forces BK to 32K translation to
be done (even if data is already 32K).
Error control same as in MSAV.
Buffer

/;,----[:=~ ~:
X REFM(A, MODE)

"'~Each

(BUFFER SPANNING NOT ALLOWED)
translate from 8K to 32K format
translate from 32K to BK format

Translates Buffer A and its associated
I.D. area according to MODE*

of the five 2048-channel data buffers has a 31 word I.D. area
associated with it. (See page 22)

Buffer

(BUFFER SPANNING NOT ALLOWED)

----Starting word
-----Number of words to operate on
(Must be <= 2048)
-----Number of places shifted (ignored i f
mode~ 0) >O shifts right;<O shifts left
<O HI LO HI LO ... HI LO -LO LO ... LO HI HI .. HI
= circular shift
>O LO LO ... LO HI HI .... Hi-HI LO HI LO ... HI LO
X EXCR(A,S,L,D, MODE)

General Purpose Exchange Routine performs
exchange on an area L words (not channels)
long beginning at word (not channel) S
in buffer A. Mode specifies type of exchange. If mode = 0, exchange is a
circular shift of D places starting at
word S in buff er A. (D > 0 to shift
right,< Oto shift left).
If mode> 0, takes L/2 low order words
followed by L/2 high order words, starting
at word S in buff er A, and pairs them
in order HI LO HI LO ... HI LO. (Similar
to X pair in 8K Focal)
If mode< 0, takes L/2 pairs of words
(HI,LO) starting at word S in buffer
A, and shuffles them into L/2 low order
words followed by L/2 high order words.
Note: If mode f 0, L must be an even
number.

21
e) IBM Tape Functions
The 9-track, IBM compatible tape transports connected to the PDP 8I
computers can be controlled using commands in the Lick 32K FOCAL language.
These commands allow reading or writing of single records of up to 2048
channels of data. The PDP 8I's use two 12-bit words per IBM 32-bit word,
ignoring the most significant 8 bits of each IBM word. Positive and negative numbers of magnitude less than 2 22 -1 are handled correctly.
Data is written as "records" or "blocks" at a density of 800 8-bit
bytes per inch, up to 8200 bytes per record (8192 bytes data, 8 bytes
control information).

Records are separated by a record gap of 0.6" or

larger of erased tape. "Tape Marks" or "End of File" marks are followed
by about 3.5" of erased tape. In order to locate a particular record on
tape, one must count records from the beginning of the tape, or one may
count "End of File" marks and then count records within a particular file.
When writing tape, a "read after write" error test is made. If an
error is detected, the tape will backspace over the bad record, erase 5"
of tape, and try again on the next segment of tape. The IBM format allows
gaps of any length greater than 0-.6", so that large segments of bad tape
can be safely skipped over by the writing program.
The following IBM tape commands are available:

Buffer (BUFFER SPANNING NOT ALLOWED)

Channel
(-1 to read 31 words into I.D. area A)

Number of Channels (ignored if C= -1)

~~------------------Starting

SET D=FREAD(A,C,N)

Reads N channels from IBM tape to buffer A
starting at channel C. Reads to end of
buffer i f N=O.
D=4096 for normal read.
D=O for end of file. Tape left positioned
past end of file.
D is negative if read parity error occurs
more than 3 times; reads anyway.

Notes:
1. If the number of channels on tape record exceeds N, excess channels
are ignored, and no error indication i~ given. These excess channels
are skipped completely; they will not be read by the next READ.
2. If the number of channels on tape record is less thant N,d?nf~ydavaanidlable
channels are read. Remaining channels in core are no mo i ie ,
no error indication is given. See LREC command for READ length.

21. 1

Buffer

(BUFFER SPANNING NOT ALLOWED)

r~-~-~-~ Starting

/ '/ '

1/'

SET D=FWRIT(A,C,N)

channel
(-1 to write 31 words from I:D. area A)

Number of channels (ignored if C= -1)
Writes N channels to IBM tape from buffer A
starting in channel C. Writes €1l end of
buffer i f N=O.
D=O if write successful on first try.
D>O gives count of number of segments of
tape that had to be erased before
write was successful.
D<O if write fails after 3 attempts. Tape
is left positioned prior to start of
last bad write attempt.

Notes:
1. If C= -1, the 31 word I.D. area is written. This I.D. area.will appear
on tape as a 16 channel record.
2. In order to detect tape slippage, FWRIT will time out if it does not
find a record gap pulse within 0.1 seconds of finishing a write. If
slippage is detected, FOCAL quits with a ?2-5151?00.00 error message.
If this happens, try cleaning tape capstan.
3. When writing many records, or whenever you are writing to a fairly
full tape, the tape status should be checked after every write (See
FIBM command below)~ When "end of tape" is sensed after a write,
two end of file marks should be written (See X EOF cotmnand below).
Then the tape should be rewound, unloaded, and a new tape mounted in
its place. If one fails to check for end of tape, in the course of a
long night of data taking, one might fill a tape and end up winding
the tape off the end of the reel, thus losing data and time.
X IBME (0) - - - - - - - - - - - - - E r a s e s 4 feet of tape.
X EOF(O)-----------

Writes an' end of file mark. Tape is
left positioned past file mark.

Notes:
1. In order to do any operation that writes on the tape, a WRITE RING
must be inserted into the back side of the tape reel hub. If one
uses any of the FOCAL commands which write on the tape (WRIT-, EOF,
or IBME) on a tape without a WRITE RING, FOCAL will start printing
PROTECT
every 2 seconds. If writing is actually desired, and i f the writing
is to conmence at the start of the tape, then one should:
a) Switch tape drive to OFF LINE~ (FOCAL now starts printing
OFF ~ instead of WRITE RINGJ
b) Rewind and unload tape.
c) Insert WRITE RING into back of tape reel hub.
d) Remount tape and position at load point.
e) Switch tape drive to ON LINE. FOCAL will stop printing
OFF LINE and will write the d~sired record at start of
tape-.- 2. One can check for a WRITE RING before attempting to write by using
the FIBM command to check the tape status.

21. 2

SET D=FIBM(M)

D is set equal to the tape status,
masked by M unless M=O.
Status bit values:

2048 - END OF TAPE
(Use this when writing)

1024 - TAPE READY
(Tape unit power is on, and tape
is tensioned, positioned on or
past load point)

512 - FILE PROTECT
(No write ring is in tape reel hub
i.e., tape cannot be written on)

256 - REWINDING
(Tape drive is busy rewinding)

128 - BEGINNING OF TAPE
(Tape is positioned at beginning,
i.e., load point or B.O.T.)
8 - ON LINE

(Tape transport is switched on line.
This is like the REMOTE position
on DECtape drives; OFFLINE is like
LOCAL)

X RWND(O) ~~~~~~~~~~~~~Rewind up to beginning of tape (BOT).
X HUNT(O)~~~~~~~~~~~~~-Used to locate end of ·data on tape (not
physical end of tape). Moves forward till
double end of file (EOF) is found and stops
following the double EOF. The teletype
will type "EOF" each time a file mark
passes the read head, so that one can use
X HUNT to count the number of files on
a tape.

21.3

SET D=FADV(N,J) ~----- Advances N records.
If J=O, stops past EOF if EOF found.
If J~O, continues advancing past EOF.
D is set= 4096*(# of file marks passed)
+ # of records with parity errors
SET D=FBAK(N,J) ------Backspaces N records.
If J=O, stops past EOF if EOF found.
Note- Two "EOF" messages are
printed: 1 for backspacing over
the EOF, and 1 for advancing
back over it. D is set to 4096
plus # of records with parity errors.
If J~O, continues backspacing past EOF.
D is set = 4096 * (# of file marks passed)
+ # of records with parity errors.

=O for record length
=-1 for record number
SET D=FLREC(MODE)

If MODE=O, D is set to the length (in channels)
of the last record read or written on IBM tape.
If MODE= -1, D is set to the record number
(modulo 4096) of the last record read or written
to IBM tape. The record number is a count of the
number of records (including EOFs) from the start
of the tape.

Notes
1. The IBM tape record number counter is set to 0 by X RWND(O)
2. File marks are counted the same as data records.
3. The IBM tape record number counter is displayed in the MQ
register while the tape is in motion during X BAK or X ADV.
4. File marks do not have a defined length, i.e., if FLREC(O)
is used immediately after reading or writing a file mark,
the length returned is unpredictable.
5. Remember that the IBM tape record counter counts modulo 4096;
thus FLREC(-1)=0 does not necessarily mean one is at the
beginning of tape.

A FINAL NOTE ON IBM TAPE COMMANDS
By adding 4096 to the first argument of the following commands, the
printing of the "EOF" messages on the teletype is suppressed:
READ
HUNT
ADV
BAK

21. 9

f) I.D. Functions
Each of the 5 2048-channel data buffers has a 31 word ID area associated with
it.

All 5 of these I.D. areas are available as long as the computer being used

has at least 12K of core memory.

In addition to these 5 I.D. areas (numbered

0-4), there is an additional 31 word I.D. area which can be referenced by using
a buffer number of -1.

All of these I.D. areas are safe over bootstrapping.

(NOTE - Words 0 to 3 of I.D. buffer -1 are used by the P/S monitor; see Appo C)
The MSAV and MGET commands (see page 19) transfer both the I.D. buffer and
its corresponding data buffer to and from the DECtape. However, as there is no
data buffer -1, I.D. buffer -1 cannot be transferred to or from DECtape using
MSAV or MGET. Neither can the IBM tape commands READ or WRIT be used with I.D.
· buffer -1.
The conversion table below shows the correspondence between the location
of the I.D. words on an SK format DECtape scan, and the 32K FOCAL I.D. buffer
locations that these I.D. words will occupy. Note that this table applies to
204S channel scans; the I.D. information from a 4096 channel scan will occupy
two separate I.D. buffers.

CONVERSION TABLE - LOCATION OF I.D. WORDS - 8K .VS. 32K FORMAT
8K WORD 1fa
121

122

123

124

125

126

127

128

SK
BLOCK
1J:

32K FOCAL I.D. Buffer Word Numbers are shown in corresponding box

** Reserved for use by 32K FOCAL

Word (0-30)
Buffer
Variable to Store

~ t
P

X IPUT (W ,A, D,P)

0 for double precision

Store single (or double) precision variable D into word(s) W (and W+l) of I.D.
area A.

Word (0-30)
f ~Buffer

,!' ~ P t 0 for double precision
Set D=FITAK(W,A,O,P)

Read single precision variable from word
W, I.D. area A (or double precision variable
from words W, W+l)
1st Word number (0-30)
Buffer

I ;1~--~~- No.

;1
X TYCO(W,A,N)

of characters (1-62) (rounded up to next
even number)
Accept up to a maximum of N characters
from the teletype and pack them 2 characters
per word starting at word W in I.D. area A.
If used in the function form (i.e., SD=
FTYCO ( ... ) ) returns the number of characters
accepted, or -1 if ESCAPE or ALTMOD was hit.

- - - - - - 1st word number (0-30)
Buffer
No. of characters (1-62) (rounded up to next
even number)
X COTY(W,A,N)

Starting at word W in I.D. area A, unpack
and print I.D. information for up to a maximum of N characters. Characters are assumed
to be in packed format, 2 characters per
word. Printing stops if a null character
is encountered. PRINTS ON EITHER TELETYPE OR CRT. If used in the function form
(i.e., S D=FCOTY( ... ) ) returns the number
of characters printed.

g) Scanner Memory Control Functions
X MEME(O)
Set D = FMEMC(N)

X MEMW(W,N, OR)

Erase scanner memory
Set counting time N cycles (~ 222 -1).
Returns remaining counting time. Stops if N=O.
Doesn't set time if already counting.
Write N channels starting at Channel 0 (N<=2048)
Buff er 0 of core memory to scanner
memory, starting at scanner channel W.
Writes Low 12 bit part if OR=O, otherwise
writes High 12 bit part.

Set D = FMEMR (A,B,H)

Read scanner memory to core. First 2048
channels read to buffer A (not read if
A = -1) second 2048 channels read to buffer
B (not read if B = -1) reads all 24 bits
of each channel if H = O; reads only Low
12 bits if H ~ 0 (much faster). Sets D
if counting (SD= F MEMR( ... ) )

X PAUS(N)

Stop counting if N = 0, continue if N = 1.
No change in actual live time.

h) Double Buffer Instructions
(require at least 20K of core)
Output buff er
Input buffer
- - - - - - - Starting channel in buffer A
Starting channel in buff er B
Number of channels
Moves N channels, starting at channel CB
in buffer B,to buffer A, starting at
channel CA. Contents of buffer B unchanged
A(x) = B(y)

For X = CA to CA+N-1
Y = CB to CB+N-1
, . - - - - - - - - - Augend register
,________ Addend register
-------Starting channel in buffer A
- - - - - - - - Starting channel in buffer B
-------Number of channels
Adds N channels, starting with channel CB
in buffer B,to buffer A, starting at
channel CA. Contents of buffer B
unchanged
A(x) = A(x) + B(y)
For X = CA to CA+N-1
Y = CB to CB+N-1
,------------ Minuend buffer
----------- Subtrahend buff er
------Starting channel in buffer A
Starting channel in buffer B
-----Number of channels
Subtracts N channels, starting with channel
CB in buffer B,from buffer A, starting
at channel CA. Contents of B unchanged
For X = CA to CA+N-1
A(x)

= A(x) - B(y)

= CB to CB+N-1

,.--~~~~---Multiplier/Product

Buffer

Mul tiplican:;I Buffer
---~~~~~Double Precision Constant Multiplicandt
,-----~~Single

Precision Divisor for scaling

-----Starting channel in Buff er A
/

~Starting

channel in Buff er B

~~Number

of channels to process
Does Double precision multiply of
N channels starting at channel CA in
buffer A and CB in buffer B. Result
goes to buffer A; Buffer B is unchanged.
Result is divided by single precision
constant D, which defaults to 1.

X DMUL(A,B,M,D,CA,CB,N)

A(X)

= (A(x) * B(y) ) / D for x=CA to CA+N-1
for y=CB to CB+n-1

. tif Mt 0, Buffer A is multiplied by constant M,
i.e., A(x) = [A(x)*M)/D] for x =CA to CA+N-1.
(B and CB are ignored in this case)

,--~-----Dividend/Quotient
..---~--~-Divisor

Buffer

r------Single precision Multiplier for scaling
~----~-Double precision constant divisor *
,------Starting channel in Buff er A
/

~ Starting

channel in Buffer B

~ ~ Number

X DIVD(A,B,M,D,CA,CB,N)

of channels to process
Does double precision divide of N channels
in buffer A by N channels in buff er B starting
at channel CA in Buff er A and in channel CB
in buffer B. Result goes to buffer A; buffer B is unchanged. Result is multiplied
by single precision constant M, which defaults to 1.
A(x) = (A(x) * M) / B(y) for x=CA to CA+N-1
for y=CB to CB+N-1

* If D f 0, Buffer A is divided by constant D, i.e.,
A(x) = [A(x) * M]/D for x=CA to CA+N-1
(B and CB are ignored in this case)
Note: Division by zero results in a value
of zero.

i) Notes
A NOTE ON ARITHMETIC FUNCTIONS

DMUL
DIVD
LOGB
PUTN

OUT

POLY
PUTN

If these commands are used in the function form (i.e., SD= FDMUL ( ... etc.) ),
D is set equal to the number of channels in which arithmetic overflow occurred.
Negative overflows are set to -(2 23 -1); positive overflows to +2 23 -1.
Notes on Double Buffer Instructions
1.

If N=O, the instruction terminates as soon as channel 2047 of either
buffer has been processed.

Channels are processed one by one, from left to right.

If N~O and N exceeds the number of channels remaining in either buffer,
the instruction terminates as soon as the last available channels in either
buffer has been processed: NO ERROR MESSAGE IS GIVEN IN 11IIS CASE.

IF 11IE AREAS OF MEMORY IMPLIED BY (A, CA, N) AND (B, CB, N) OVERLAP,
RESULTS MAY NOT BE THOSE DESIRED: NO ERROR MESSAGE IS GIVEN IF OVERLAP
OCCURS:
Overlap cannot occur if N=O and A;fB.
Overlap is sometimes desirable; for example, a quick way to multiply a
buffer by 2 is to add it to. itself. i.e .• X INCA.A). Care should be
exercised when specifying arguments that cause overlap to occur.

26. 1

Miscellaneous commands
SET D=FLO (X)

Return as an unsigned 12-bit
integer the low-order 12 bits
of 24-bit signed integer X.

SET D=FHI (X)

Return as an unsigned 12-bit
integer the high-order 12 bits
of 24-bit signed integer X.

SET D=FAND(X,Y)

Returns as a signed 24-bit integer
the logical AND of the two 24-bit
signed integers X and Y.

SET D=FBIG (X)

Return as a 24-bit signed
integer the 12-bit signed
integer in X. (The 12- bit
number X is copied into the
low order 12 bits of D; then
the sign bit of X is extended
into the high order 12-bits of D.)

III.

32K FOCAL OVERLAY FUNCTIONS
a) Description
The commands that have been described to this point are the ones most

frequently used and as such have been made a permanent part of 32K FOCAL.
Additional commands that are used less frequently can be added to the program
DECtape as "overlays" using the system program TOVR (see LOTR 32K Focal System
Description).

The commands in these overlays can then be made available to a

Focal program by using the X NAME command (see pg.

12 ).

An "overlay" can contain from 1 to 10 Focal functions, depending on

the length of each function.

All ofthe functions residing in the same over-

lay become available whenever that overlay is loaded into either of the 2 NAME
buffers.

All of the overlays described on the following pages can be used

in either of the 2 NAME buffers.
However

overlays operating

on the data buffers will run slightly faster in NAME buffer 2, and overlays
using the Serial Data Multiplexer (see LOTR 16) will run slightly faster in
NAME buffer 1.
In the future, additional overlays are likely to be developed and described in other technical reports.
b) Polynomial Overlay

r~~~~~~~~~~

Position of coefficient (O<= POS<= 10)

,.-~~~~~~~coefficient

X SAV(POS; COEFF)

Sets coefficients for X POLY.
Note: No X NAME commands should be
given in between setting coefficients
with X SAV and using them with X POLY
because the coefficients are stored in
the overlay area with X SAV and POLY!
NOTE SEMICOLON BEFORE COEFF.

Buffer
- - - - - - Order of Polynomial
,..-~~~~~ Starting channel

(O<=

ORDER<=lO)

------Number of channels
Generates a polynomial of order "ORDER"
on
contents of buffer A, starting at
channel C and processing N channels.
Coefficients are set using the X SAV
conunand. Note: no X NAME conunands
should be given and in between setting
coefficients with X SAV and using them
with X POLY, because the coefficients
are stored inthe overlay area with X SAV
and X POLY!

c) Dead.tiue, Extinction, Logarithm

Overlay

,______ Data buff er
,____ K-lambda buffer
----- 256* air mass
starting channel in A
; - starting channel in B
/ ;-Number of channels
SET D=FTINC(A,B,Z,CA,CB,N)
Extinction correction. Expects a table
of 4096*1), in buffer B (~ = extinction
coefficient for each channel). Data in
buffer A is multiplied by exp(~Z).
D is set to number of times overflow occurred.
Data buff er
Dwell Time (seconds)
l/(131072*deadtime) seconds

/;---

SET Q=FTIM(A,T,D,C,N)

Starting channel
Number of channels
Deadtime correction finds number of real
counts from observed. For each channel,
computes
x

--)
NR = counts* (1 +
1-x

where x = (counts*4096/T)*D
Q is set to number of times overflow
occurred.

A(X) = lOOOO*LOG lO(A(X))

d) Peak Finding and Linearization Overlay
------Buffer
r----- Peak channel
,.-----Continuum
- - - Threshold
Mininrum peak width (<4096)
Starting channel
of channels
SET Dl = FIND(A,XP,CT,CD,MW,C,N)
1st moment peak finder. This routine
SET D2 = FSIG(O) - Dummy argument
finds an accurate position for a peak
whose approximate position is already
known (see LOTR #12 pg. 14).

12
~Number

Returns:
D2 = :. FCHAN (J +C)
J

Dl = ~ J* FCHAN(J+C)
J

or Dl < 0 if error occurs
(overflow or peak too narrow)

FOCAL Routine to Use FIND/FSIG
1.10 SET PK=FIND(A, ••• ,C,N); IF (PK)l.2; SET PK=(PK/FSIG(O))+C; GO 1.3
l.20 C - error recovery
1.30 C - continue
Note that

FIND returns number of peak position relative to channel c.
FSIG

....------- Input buff er
.-------Threshold (Double precision)
,,..._____ Replacement flag
,-----Replacement value (double precision)
Starting channel in A
Number of channels in A
;' ~utpu~ buff er
~~Starting channel in B
SET D= FSEEK(A,TH,RF,RV, CA,N,B,CB)
Course peak finder and wild point
destroyer searches for peaks in buffer
A starting at channel CA and searching
N channels. (A peak is a series of consecutive channels all of whose values are
>=to the threshold TH). The mean
position of eac~ peak (truncated to
the nearest integer channel number) is
stored in buffer B. The position of the
Nth peak is stored.in channel CB+N of
buffer B. The position recorded for the
peak is a point halfway between where the
peak sticks up through the threshold and
where it falls back below it again .
. The number of peaks
found is returned as D and is also
stored in channel CB of Buffer B. If
RF~O, any point >TH is replaced with RV 1 .
Bin size table buffer and scrunched
data output buffer
Channel offset
Unscrunched data input buffer
Starting channel in Buffer A
Number of Scrunched channels
Starting channel in Buff er B
SET D = FSCRN(A,W,B,CA,N,CB)

Scrunches data from buffer B to
a table of bin sizes in Buffer A.
W specifies the position in the unscrunched data of the ~ft edge of bin
0, times 4096. For instance, the
middle of channel 1 has W = 6144
(l.5*4096). The bin size table contains the number of unscrunched channels, times 4096, in each scrunched
channel. Bin sizes must be greater
than or equal to 0.

e) Sweep,Time, Position Overlay (Uses Serial Multiplexer)
0 for Mux unit 3; 1 for Mux unit 4(coude)
X MCEN(C,0,0,M)

Offsets sweep center by C (+, -)

,.--~~~~~~~~~o

to suppress ramp; 1 for normal sweep

~------~~----~~sweep

center (+,-)

1 ,/

0 to write; 1 to read

,/ J

0 for MUx unit 3; 1 for Mux unit 4

X MEMX(N,C,R,M)

Loads X sweep

~----~----~~~~-Sweep

//"
~/,,/'
X MEMY ( O, C, R, M)

center (+,-)

0 to write; 1 to read
0 for Mux unit 3; 1 for Mux unit 4
Loads Y sweep

5-PST calendar
4-UT calendar
3-Sidereal calendar
2-PST clock
1-UT clock
0-Sideral clock
SET D=FTIME(C)

Returns time in seconds or date coded
3600*month + 60*day + year where Year
is the last 2 digits of the year.
<O for HA in seconds of time
{
~~~~-=O for RA in l/lO's of seconds of time

/
SET D=FPOSN(C)

>o for DEC in seconds of arc
Reads telescope position

f) Cross correlation overlay
- - - - Data buffer
------- Smoothing function buffer

SET Q=FCORL(A,B,CA,NC,CB,NS,D)

Convolves data in buffer A with
smoothing function in buffer B.

Computes:

A(J)=(A(J)*B(CB)+A(J+l)*B(CB+l)+ ...
... +A(J+NS-l)*B(CB+NSl))/D; for J=CA,CA+NC-NS
A(J)=O
for J=CA+NC-NS+l,NC
Notes
1. Number of channels to process (NC) MUST BE NON-ZERO.
NC=O will give an error message. Use NC=2048 to process
a full 2048 channel buffer.
2. Length of smoothing function (NS) MUST BE NON-ZERO.
NS=O will give an error message.
3. NS must be less than or equal to NC.
4. Single precision divisor D defaults to 1.
5. Result Q is set to number of times overflow occurred.

-------Buffer
I / - - - - - - - Starting channel

//------Number of channels
X FLIP(A,C,N)
Reverse content of channels
A(C) = A(C+N-1)
A(C+l) = A(C+N-2)
A(C+N-1) = A(C)

g) Debugging utility overlay

Core address
,/" ~~----------------- Verify value
~ ~,.---------------~ Replacement

SET D=FZAP(FLD,LOC,VER,REP)

value

Inspects, verifies, and replaces
specified location in core memory.
Returns in decimal-coded octal the
contents of location LOC in field FLD.
FLD, I.DC, VER, and REP are all specified
in decimal-coded octal.
If either VER or REP is non-zero,
replaces the contents of LOC in field
FLD with the value REP, provided that
the value of VER identically matches
the current contents of LOC. If VER
does not match, location LOC is not
changed, and an error message is
printed.
If FLD=LOC=VER=REP=O, returns in
decimal-coded octal the contents of
the next sequential location. This
feature is useful for core dumps, and
can spill across field boundaries.

SET D=FADDR(O)------------------------------~Returns in decimal-coded octal the
core address of the last location
referenced with ZAP.
SET D=FIELD(O)------------------------------~Returns in decimal-coded octal the
memory field of the last location
referenced with ZAP.
SET D=FDEC(X)------------------~----------~Returns decimal value of decimal-coded
octal number X. TYPE FDEC(7777) would
print 4095 on teletype. Only accepts
numbers from 0-7777.
Returns in decimal-coded octal the
value of decimal number D. TYPE FOCT(409
would print 7777 on teletype. Only
accepts numbers from 0-4095.

------------~Function

word

word
/'"-----------oData Public,
1
Private
=

SET D=FMPX(FUNC,DATA,BUS)

Transmits function word FUNC and
data word DATA to serial multiplexer.
FUNC and DATA are both specified in
decimal-coded octal. Returns data
received in decimal-coded octal •
SET D=FMPX(650) would read data from
Cable 8, Unit 5.
X MPX(4320,20) would send Pulse 1 to
Cables 8 and 18, Unit 3.
See LOTR #16, page 5.
--------Instruction field
r--------Data field
,...-------Entry point address
--------AC contents
- - - - - - - MQ contents
- - - - - - L i n k value

SET D=FJMS(IF,DF,EPAD,AC,MQjLlNK)

Does a PDP-81 JMS machine command
to location EPAD in field IF.
Immediately prior to executing the JMS
the data field is set to DF, the AC to
AC, the MQ to MQ, and the link to LINK.
IF and DF must be in the range 0 to 7.
EPAD, AC, and MQ are all specified in
decimal-coded octal. LINK can only be
0 or 1. Result D is set to 24-bit
contents of AC and MQ upon eventual
return (if ever) to FOCAL.

- - - - - - - - - Instruction field
- - - - - - - - D a t a field
- - - - - - - E n t r y point address
- - - - - - - A C contents
- - - - - - - MQ contents
- - - - - - L i n k value
SET D=FJMP(IF ,DF ,EPAD,AC,MQ, LINK)

Does a PDP-81 ·JMP machine command
to location EPAD in field IF.
DF, AC, MQ, and LINK are set as
described in FJMS instruction above.
It is unlikely that control will ever
return to this command.

h) Intensity map overlay

r--~~~~~~~~~~Data

buffer

origin
,--~~~~~~~~-Y origin

--~~~~~~~~~x

Spacing between lines (X)
Counts per dot
,--~~~~~~~ Starting channel
,--~~~~~~-Number

X LOOK(A,X,Y,ST,SC,C,N)

of channels

Draw an intensity map. Uses low 12 bits
of channel, only. Plots N vertical lines
of varying density, starting at X,Y, with
X spacing ST, intensity scale factor SC,
starting at channel C in buffer A.

IV. 32K FOCAL UPDATE - June 20, 1977
Significant changes have been made to the 32K Focal System.

The current

version of 32K FOCAL (version 77B) is described in Lick Observatory Technical
Report 1121, "32K FOCAL User's Guide" (Feb. '77).

Effective June 20, 1977,

version 77B of 32K FOCAL was replaced by version 77C*.

This update describes

the differences between the two versions, and the updates to be made to
LOTR #21 so that it will correspond to version 77C.
The differences between version 77B and 77C fall into the following areas:
I.

Modifications/corrections of resident functions

II.

Corrections of overlay functions

III.

Addition of new resident functions

IV.

Addition of new overlay functions

Relocation of resident functions

VI.

Correction of inherited bugs.

Robert Kibrick

The version number of any 32K FOCAL System DECtape or floppy disk appears
in the title line that is printed whenever a FOCAL WRITE command is
executed.

MODIFICATIONS/CORRECTIONS OF RESIDENT FUNCTIONS
A.

Corrections
1.

SET D = FMEMR (A,B,H)

(See p. 23, LOTR #21)

The "read only low 12 bits" option (H :f: 0) now works correctly.
In version 77B, use of this option would give correct results,
but would actually run slower than if all 24 bits were read.
2.

SET D = FCHEK (O)

(See p. C-3, LOTR #21)

Counting time display now counts dovm. correctly to 0.

In version

77B, counting time display would count down past 0 and leave
a final display of 4095.

Counting time now counts down in

minutes/seconds; in version 77B countdown was in seconds only.
B.

Modifications
1.

X CALL (N,S,Q)
a)

(See pgs. 12-13, LOTR #21)

Calls can no longer be nested to 10 levels as could be
done in version 77B.

Attempts to use the nesting option

(Q :f: 0) now generate a ?5422?00.00 error message.

Nested

calls can now be done using the new CALL/bo function
(See X CADO command, pg. U7

of this update)

Changes to text of LOTR #21
1)

p. 12, delete these 2 sentences under X CALL
"If Q = 1, calls can be nested to 10 levels.
Nesting cleared for Q = O."

p. B2, insert error code ?5422?00.00
"5422 X CALL NESTING NO LONGER SUPPORTED.
INSTEAD."

p. Bl, delete error code ?22.24.

USE X CADO

Using the X CALL (or X END) command from inside a DO group
or FOR loop has always been illegal (see LOTR #1, pg. Z9,
paragraph C).

However, such illegal uses were never detected

in either SK FOCAL or 32K FOCAL version 77B, and would
cause unpredictable results.

In version 77C, such illegal

uses of X CALL or X END are detected, and a ?5440?00.00
error message is printed.
Changes to text of LOTR #21
1)

p. B2, insert error code ?5440?00.00:
"5440 X CALL or X END FROM WITHIN A DO GROUP OR FOR LOOP"

In both 8K and 32K FOCAL version 77B, one could easily wipe
out many minutes worth of keypunching FOCAL text if one
accidentally started a new (or modified) program that
contained an X CALL (or X END) command, before one had
filed the new program on DECtape or floppy disk.

In version

77C:

If FOCAL's text area contains new or modified text that
has not yet been filed, then execution of X CALL (or X END)
will cause FOCAL to type
REALLY?

on the teletype, and then to wait for a reply.

Any reply

other than 'Y' will cause the X CALL (or X END) command
to be supressed, and the contents of the text area to be
preserved.

A reply of 'Y' will cause the X CALL (or X END)

to proceed, and the contents of the text area to be lost.

Changes to text of LOTR #21:
1)

Cut off bottom half of this page; paste on pg._ 12_. LOTR tnl
over

X CALL (N,S,Q)

X END (0)

(See pg. 11, LOTR #21)

Modifications b) and c) as described under X CALL above apply
also to X END.
3.

X FILE (N)

(See pg. 13, LOTR #21)

The contents of FOCAL's text area is written to DECtape or floppy
disk as program N; then the program N just written out is read
back into FOCAL's text area, as if an X CALL (N) had been used.
In version 77B, the text area is written out but not read back.
The user should not notice any real difference between the two
versions, except that the version 77C X FILE command will take
slightly longer.

Place over bottom half of page 12, LOTR #21
Cut here

------------------------------------------------------------------------------X CALL(N,S)
(For greater versatility in
calling, see the X CADO
command, pg. U7)

• Call Program N from Dec tape. If S> O,
starts program N at subroutine s.
(Code 8*128 + L to start program N at
line L of subroutine S). Use X END(O)
to return to the line following the
X CALL command in the calling program.

Important Notes:
A) It is illegal to use either X CALL or X END from inside a DO group
or FOR loop.
B) If FOCAL's text area contains new or modified text that has not yet
been filed (See X FILE command below), then execution of X CALL or
X END will cause FOCAL to type "RF.ALLY?" on the teletype, and then
to wait for a reply. Any reply other than "Y" will cause the X CALL
or X END command to be supressed, and the contents of the text area
to be preserved. A reply of ''Y" will cause the connnand to proceed
and the contents of the text area to be lost.
C) If X CALL is used to address a program area on the Dectape that does
not contain a FOCAL program, the following action is taken:

II.

CORRECTIONS TO OVERLAY FUNCTIONS
A.

Sweep, Time, Position Overlay

(See p. 32, LOTR #21)

The commands in this overlay have been changed so that they work
correctly in either NAME buffer.

This new overlay (dated 5/16/77)

will not work if used with version 77B FOCAL.

The old Sweep, Time,

Position overlay (dated 2/14/77) will not work with version 77C
FOCAL.

The old Sweep, Time, Position overlay worked correctly if

used in NAME buffer 1, but caused occasional MUX errors if used in
NAME buffer 2.

When copying version 77C to your FOCAL tape, be

sure to copy an updated version of this overlay.
Changes to text of LOTR #21
p. B3:
B.

Change error code 6171 to 6165, 6174 to 6170.

DebuggingUtility Overlay
1.

(See p. 34, LOTR #21)

The X MPX counnand in this overlay has been changed so that it
works correctly in either NAME buffer.

(The X MPX command had

the same problem as the connnands in the Sweep, Time, Position
overlay, and the same considerations apply.

The new Debugging

Utility Overlay is dated 5/4/77; the old one was dated 2/23/77).
2.

The X MPX command no longer hangs up the spectrograph control
panel at the 120" readout room; the previous version did.

·u6

III. ADDITION OF NEW RESIDENT FUNCTIONS
A. X BRK(O)
One of the problems with FOCAL is that:
"If a GOTO or IF conmand that is inside a DO group transfers control
to a line outside the DO group, that line is executed and control
then returns to the connnand following the DO." (see DEC. FOCAL manual)
A similar problem exists with GOTO or IF statements that are used inside
of FOR loops. This feature of FOCAL makes it awkward to bail out of a
DO group or FOR loop, as is illustrated by the following example program.
*01. 10 FOR J=l,3;DO 2
'*01. 20 TYPE ! "L 1. 2"
*O 1. 30 TYPE ~ "L 1. 3" ;QUIT
*02. 10 TYPE %2,J;DO 3
*02.20 TYPE !"L 2.2"
*03.10 IF (J-2)3,3,3.2,3.3
*03. 20 TYPE l "L 3. 2"
*03. 25 GO L 2
*03. 30 TYPE t ''L 3·. 3"
*GO
1
L 3.3
2
L 2.2
L 3. 2·
L 1. 2
3
L 2.2
L 3. 3
L 2. 2
L 1. 2
L 1. 3*

The X BRK connnand solves this problem
by purging FOCAL's pushdown list, which
is where FOCAL stores information
relating to the nesting of DO groups
and the control of FOR loops.
The effect of the X BRK command is to
break one out of the inside of any DO
groups or FOR loops, as illustrated by
the example below. Note that any commands
on the same line following an X BRK
command are ignored, and execution continues with the next sequential line.
*01.10 FOR J=l,3;DO 2
*01. 20 TYPE t "L 1. 2"
*01. 30 TYPE l "L 1. 3" ;QUIT
*02. 10 TYPE %2,J;DO 3
*02.20 TYPE ~"L 2.2"
*03.10 IF (J-2)3.3,3.2,3.3
*03. 20 TYPE t "L 3. 2" ;X BRK(O) ;TYPE "ABC"
*03. 25 TYPE ! "L 3. 25" ;GO 1. 2
*03. 30 TYPE !-"L 3. 3"
*GO

1
L 3. 3

L 2.2
L 3. 2

L 3.25
L 1.2
L 1. 3*

Program
/ ( ; Group
Step
B.

X CADO (P,G,S)
This command functions exactly like a FOCAL DO command, except it
allows one to "DO" text that is in another FOCAL program.

Program P

is called, then FOCAL 'DO'es group G, step S, as if it were part of
the calling program.

If S = 0, all of group G is "done."

G = S = 0, all of program P is "done".

When the specified text has

been "done," control returns to the statement following X CADO in
the calling program.

The X CADO command can be used anywhere a

FOCAL DO command can be used, and X CADO's can be nested.

The degree

of nesting allowed depends upon space available in the pushdown
list; X CADO requires l more word of pushdown list space than does a
regular FOCAL DO statement.
Note - If FOCAL's text area contains new or modified text that has
not yet been filed, then execution of an X CADO connnand will
result in a ?3336?00.00 error.
Changes to text of LOTR #21
p. B2, insert error code 3336
"3336 X CADO EXECUTED FROM MODIFIED TEXT BUFFER.
FIRST."

USE X FILE

U7.l

--..ouffer
Starting channel

~Number of channels
C.

X INV(A,C,N)
Negative of buffer
A(X) = -A(X); for X = C to C + N - 1

ouble Precision increment
/{'Starting channel

j r-Num.ber of channels
D.

X INCR (A, I , C, N)
Increment a buff er by a constant
A(X) = A(X} + I; for x = c to c + N - 1
FINCR(. •• ) re.turns overflow count •

.-----Starting word in A
,.....~~~~-Output

buffer

,.--~~~~~umber

of words to copy

E.
Copies NW words of I.D. information from I.D. buffer B starting at
word WB to I.D. buffer A starting at word WA.

2 new functions, X PV and GV allow one to store 18 single precision
numbers into an area of memory that is safe over both power failures

U7.2

and re-bootstrapping on any of our PDP-8s.
A location (0-17)

single precision number (0-4095)

X PV(LOC,VAL)
Stores value into indicated location.

location (0-17)

S D = FGV(LOC)
Sets D equal to value of location LOC.

Occasionally, one needs to execute a sequence of FOCAL commands
without interruption.

The X IOF

command allows one to prevent

FOCAL from responding to the CONTROL/C key, while leaving the
computer's interrupt enabled.
X

IOF(O)

Causes FOCAL to ignore any subsequent CONTROL/C's
until the next X IOF (-1) command.

X IOF(-1)

Causes FOCAL to respond to any subsequent CONTROL/C's
until the next X IOF(O) command.

If one has mistakenly set X IOF(O) and needs to abort the FOCAL
program, hit the STOP button on the CPU.

Then set switch registers

to octal 0200, hit load address, and START.

This implicitly

executes an X IOF(-1) and executes the same restart sequence as
if FOCAL had been aborted with CONTROL/C.

One need not re-bootstrap.

IV.

ADDITION OF NEW OVERLAY FUNCTIONS
A.

Quick Data Overlay
- - Buffer (BUFFER SPANNING NOT ALLOWED)

,__

Starting channel
(-1 to read 31 words into I.D. area A)
of channels (ignored if C = -1)

SET D =
Starts reading N channels from IBM tape to buffer A.
of buffer if N = O.
1)

Reads to end

Works like READ command except:

Control returns to FOCAL as soon as the IBM tape starts moving.
Data transfer can go on in parallel with computing.

(One must be

careful not to use the portion of buffer A into which the data is
being transferred until the transfer is completed!

Use DONE*

command to check for completion.)
2)

D returns the status of any previous IBM read or write operation,
not the status of this command.

See DONE command for description

of status value returned.
3)

If the tape drive is still busy processing a previous RDQ or
WRQ command when the current command is executed, three options
are available

w= 0

A ?6062?00.00 error is printed and the FOCAL
program

w< 0

is terminated

The current command is ignored and control
returns immediately to the FOCAL program.
is set -1

>'<'The DONE command is described on page UlO

w> 0

The current command waits for the previous IBM
tape operation to complete.

Then the current

command is carried out.
4)

If the IBM tape is at END OF TAPE (EOT), the tape drive will
appear "busy".

The same actions are possible as described

in 3) above, except that in the case of W < 0, D will be set
to -2.

Buffer (BUFFER SPANNING NOT ALLOWED)
Starting channel
,.--- (-1 to write 31 words from I. D. area A)
(ignored if C = -1)

SET D = FWRQ
Starts writing N channels to IBM tape from buffer A, starting in
channel C.

Writes to end of buffer if N = O.

Works like WRIT

command except for the same 4 differences described under the

RDQ command above.

Also, if a WRQ command is directed to a write

protected IBM tape, a ?6273?00.00 error is given.
Notes:

The RDQ and WRQ commands can be used interchangeably with
READ and WRIT, and are fully compatible with all other
resident IBM tape commands, with the exception that one
should allow approximately a 20 millisec delay after using
X BAK before using X RDQ or X WRQ.

Changes to text of LOTR #21:
p. B3 - add error codes 6062 and 6273
"6062

RDQ or WRQ USING 'NO WAIT' OPTION FOUND IBM TAPE BUSY

6273 WRQ FOUND IBM TAPE WITHOUT WRITE RING"

UlO

SET D = FDONE(O)
Returns status of last READ or WRITE operation.
D = 0 if last RDQ or WRQ operation completed successfully
D < 0 if last RDQ or WRQ operation has not yet completed
D > 0 if tape error on last RDQ or WRQ, or if last record was a
file mark.
The low 12 bits of D contain the remaining word count
The high 12 bits of D contain:
BIT 11 is 1 if last record was file mark

(4096 bit)

BIT 10 is 1 is parity error on last record

(8192 bit)

BITS 1-9 unused
BIT 0 is 0

,--~~~-Number

to be queued.

(0 .:s_ A .:s_ 4095)

SET D = FADDQ (A)
Appends A to the end of a 5 element queue.*
D = 0 if successful.
If queue is already full, D is set < 0 and queue remains unchanged.

SET D = FDLTQ (O)
Sets D to the value of the front element of a 5 element queue*,
then removes this element from the queue.
Sets D < 0 if queue already empty.

Notes:

ADDQ

and DLTQ are useful for maintaining a queue of

buffers to be read or written using the RDQ and WRQ commands.
By queueing data, one can handle instantaneous data rates

* A queue works on the basis of first in, first out.

Ull

that exceed that average data transfer rate of the IBM tape
drive.

Such situations occur in high speed data taking

applications.
2.

The storage area used for the queue is contained within the
overlay.

The queue is initially empty when the overlay is

brought in with X NAME.

No X NAME commands should address

the buffer containing this overlay in between saving queue
elements with ADDQ and retrieving them with DLTQ.

~Number of scanner cycles to count

/ S w i t c h mask

SET D = FMCQ (N, MASK)
If scanner is already counting or if N < 0, returns remaining
counting time, and leaves scanner unchanged.

If scanner is not

counting and N ~ 0, then
1)

Waits for a pulse to be detected on the switch(es) in group 3
selected by MASK.

(External clock pulses can be brought in

across switches 3.8, 3.9, and 3.10.

Pulse width should be at

least 30 µsec)
2)

Waits to synchronize with scanner memory (max. 4.4 millisec).
D returns time spent waiting to synchronize with memory in
units of 13.25 µsec.

Once synchronized, sets counting time of N scanner cycles and
starts counting.

1 scanner cycle is ::::: 4.4 µsec.

a counting time of ::::: 10 hours).

(N = 0 gives

Ul2

~~~~~~~~Buffer
~;1~~~~~~-Buffer

for 1st 2048 channels (right slit)
for 2nd 2048 channels (left slit)

X MRQ (A,B)
Wait for current

scan to complete, then read scanner memory to

core.
First 2048 channels read to buffer A, (starting in channel 0)
and erased from scanner memory as they are read (Not read or
erased if A= -1).

Similarly for 2nd 2048 channels and buffer B.

Reads all 24 bits normally.
To read only the low 12 bits of each scanner channel, and to pack
2 of these 12 bit scanner channels per 24 bit buffer channel, add
4096 to the buffer number.

Data Compaction Overlay
Output Buffer
Input Buff er
Starting word in A (0 .::. WDA _:::. 4095)
~~Starting word
/~Number

X PAK (A,B,WDA,WDB,N)

in B (0 .::. WDB .::. 4095)

of channels in B

Packs N channels into N words

Each channel in B consists of 2 12-bit words.

Starting at WDB in

buffer B and processing N such channels, the low order word of each
channel is stored into consecutive words in buffer A, starting at
WDA.

The high order words of each channel in B are ignored.

Ul3

Output Buffer
Input Buffer
Starting word in A

Starting word in B

//Number of channels in.A:'
X UNPK (A,B,WDA,tIDB,N)
Reverse of PAK.
Starting at WDB in buffer B and processing N words, appends a high
order word of 0 to each word and stores into successive channels
in buffer A, starting at WDA.
Changes to text of LOTR #21:
p. B3 - add error code 6255
"6255 starting word(s) <O or >4095 in PAK or UNPK"

Isophote Display Overlay

(Insert as pg. 38 of LOTR #21)

Buffer
~/Starting

channel

//Number of channels
SET D =·FIPHO (A,C,N)
Takes data values in the range 0, 1, or 2 and displays them as
PIXELS of 3 different densities on CRT.
0 displays as a blank PIXEL.
1 displays as a lined PIXEL.
2 displays as a solid PIXEL.
Display consumes 1 channel of data per PIXEL, sweeping the CRT in
a raster pattern.

PIXEL size and CRT screen parameters are set

using the X PIXL connnand.
by this command.

D is set to number of PIXELS displayed

Ul4

Changes to text of LOTR #21:
p. B3 insert error code 6344.
"6344 screen full in XIPHO or XPIXL not called first."

PIXEL size (CRT dots on a side)
PIXELS per line
Lines per screen
X origin for picture on CRT
Y origin for picture on CRT
~Oto

raster from bottom to top

~ 1 to raster from top to bottom

X PIXL (PS,PL,LS,X0,Y0,YSWEP)
Defines PIXEL size and screen parameters used by IPHO command.
The input arguments must pass the following tests:
1)

PS must be 2 or a multiple of 4

PL, LS, X0, Y0 must all be > 0

PL*PS+X0 must be < 1024

LS*PS+Y0 must be < 1024

Changes to LOTR #21:
p. B3 insert error codes 6301, 6302, 6303
"6301 Bad PIXEL size in X PIXL
6302 Bad combination of PL, PS, X0 in X PIXL
6303 Bad combination of LS, PS, Y0 in X PIXL"

Ul5

Data Buffer
Translate Buffer
Length of translate buffer
Starting channel in B
Starting channel in A
Number of channels to translate
//Default low (Dbl Pree.)

Default high (Dbl. Pree.)

X TRAN (A,B,NSB,CB,CA,N,TL,TH)
This command allows one to apply an arbitrarily defined function to a
buffer of data. The function values are assumed to have been previously
computed and stored into the translate buffer. The function value stored
into a given channel of the translate buffer is simply the value obtained
by applying the function to the number of the channel. Specifically, the
TRAN co'lllDB.nd operates as follows:
Translates buffer A, starting at channel CA and translating N
channels.
Buffer B contains the translate table, starting in channel CB and
running NSB channels long.

The following operation is performed

on each indicated channel:
The value of the channel in buff er A is used as an index into
the translate table in buffer B.

If this index value is >O

and <NSB, the value of the indexed channel in buffer B is used
to replace the value of the indexing channel in buffer A.

the value of the indexing channel in buffer A is <O, then it is
replaced with the double precision value TL.
>

If the value is

= NSB, then it is replaced with the double precision value TH.

·u16

RELOCATION OF RESIDENT FUNCTIONS
The starting addresses of the following resident functions have been
moved.

In some cases, this has caused a corresponding change in the

error messages produced by these functions.

Appendices Al-A3, and

B2 should be updated accordingly.
NAME

OLD ADDR.

NEW ADDR.

ADV
AND

5257
3730

5256
4770

CHEK

3722

3732

END

5537

5550

FILE

5473

5501

PAUS

3630

3627

RWND
VAR

5121
5260

5120
5255

OLD ERROR

NEW ERROR

5273

5270

5277

5274

Ul7

VI.

CORRECTION OF INHERITED BUGS
A.

If one used a DO statement to invoke a single line that contained
an error, FOCAL diagnosed the error as occuring on the line containing the DO, and not on the line in which the error actually
occurred.
Example

*1. 2

DO 2.1

*1. 3

>t2 .1

TYPE A/r/J

*G
?28.73@ 1.20
B.

error should be diagnosed on line 2.1

The Lick FOCAL functions X DO and X GO did not behave the same as
the standard FOCAL DO and GO commands.

In particular, these commands

had the same effect as the new X BRK connnand.

That is, X DO and

X GO would purge FOCAL's pushdown list and cause the program to
break out of any DO groups or FOR loops in which it was nested.
Also, any text following X DO on the same line was ignored.

32K FOCAL version 77C, X DO and X GO now work exactly like FOCAL's
DO and GO commands.
C.

If the X EOF(O) command was immediately followed by an X BAK(l 1 1)
command, one would fail to backspace behind the file mark. Now fixed.

If an IBM tape were backspaced to beginning of tape (BOT) using the
X BAK command, or rewound using the X RWND command, the IBM tape
drive was erroneously left in the "start" state.

If another IBM

Ul8

tape command immediately followed, it could possibly start too
soon and cause I/O errors. This problem fixed in version 77C.

APPENDIX Bl - ERROR DIAGNOSTICS OF FOCAL :1.969
CODE
CONTROL-C FROM KEYBOARD OR MANUAL START
ILLEGAL STEP OR LINE NUMBER USED
GF::oup NUMf:EF;: I'.:; TOO L.(:1F;:GE
DOUBLE PERIODS FOUND IN A LINE NUMBER
?():I.{•; ~5
LINE NUMBER IS TOO LARGE
"?():l~!i4
GROUP ZERO IS AN ILLEGAL LINE NUMBER
NONEXISTANT GROUP REFERENCED BY 'DO'
NONEXISTANT LINE REFERENCED BY 'DO'
?()2 ?9
STORAGE WAS FILLED BY PUSH-DOWN LIST
NONEXISTANT LINE USED AFTER 'GOTO' OR 'IF'
ILLEGAL COMMAND USED
LEFT OF "=" IN ERROR IN 'FOR' OR 'SET'
E><CE '.:;s 1:< I GHT TE F<M I N(:1T01:;;~:; ENCCJUr··~ TEF~E:.D
'i'()4 ., (~,()
ILLEGAL TERMINATOR IN 'FOR' COMMAND
? () .:'.} t. :.":)
LINE TOO L.ONG
?O~.::i*tl
PROGl:;;1'~M TOO L.OND
BAD ARGUMENT TO 'MODIFY'
IL.LEGAL USE OF FUNCTION OR NUMBER
?Oci. ~~i4
STORAGE IS FILLED BY VARIABLES
T 0 ·.:.~· .:- ~.:.~ ~. :~
OPERATOR MISSING IN EXPRESSION OR DOUBLE 'E'
?0'7 ~ :·5B
NO OPERATOR USED BEFORE PARENTHESIS
?0?.:7 (?07.:9) NO ARGUMENT GIVEN AFTER FUNCTION CALL
ILLEGAL FUNCTION NAME OR DOUBLE OPERATORS USED
PARENTHESIS DO NOT MATCH
'!'O?. l :I.
BAD ARGUMENT IN 'ERASE'
STORAGE WAS FILL.ED BY TEXT
LOGARITHM OF ZERO REQUESTED
'IF' ARGUMENT NOT ENCLOSED IN (), [J, OR<>
LITERAL NUMBER IS TOD LARGE
~ POWER IS TOO LARGE OR NEGATIVE
DIVISION BY ZERO REQUESTED
IMAGINARY SQUARE ROOTS REQUIRED
EXCESSIVE RIGHT ·rERMINATORS IN 'TYPE'
ILLEGAL CHARACTER, UNAVAILABLE COMMAND, OR
UNAVAILABLE FUNCTION USE.
ATTEMPT TO USE FOCAL 'I...' COMMAND CTAPE COPIER)
AFTER TOO MUCH TEXT ENTERED
?()(). 00

0 ::2 ~:5 ()

o::} :r. <:·:i

() :.:) .4 ()
03'.'.'.il

0440
0464
O'S:l.7
O.::i0'.'5
(),:'):::)4

:1.047
:1.0.;:,4
1074
11.4?
13.<:\'.'5

:1.2,so

:l.40(S

:l.753
2 0~5 7
2:? :l. 3
2~.5~51

5042
'.'.'i :I. 7;:=j
~i,S44

7:1.:l.l
/40~.5
/~.~_;4 :I.

TO :l. • 4()

IF PUSH DOWN LIST OVERFLOWS, TRY 'TYPE $' TO SEE IF UNWANTED VARIABLES
ARE PRESENT. USE 'ERASE', TO REMOVE VARIABLES, BUT RETAIN PROGRAM.

APPENDIX B2 - 32K FOCAL RESIDENT FUNCTION ERROR DIAGNOSTICS
THE FOLLOWING ERROR DIAGNOSTICS PRINT OUT AS ?NNNN?00.00
CODE

ME(IN I NG

03 17

I !... L. E G(:11... F: EC UF<: ~:; I Z E U~:; E 0 F 1'~ I... I CI< F 0 Ct1 L FUN CT I 0 N
ATTEMPT TO PASS A NUMBER > IN MAGNITUDE THAN 2~23-1 ()8388607)
ATTEMPT TO USE TAPE COPIER OPTION 1 ON PDPS WITH <12K OF CORE
X CADO EXECUTED FROM MODIFIED TEXT BUFFER. USE X FILE FIRST
FLOATING POINT ARGUMENT TO LICK FOCAL FUNCTION NOT LAST IN LIS'
SEMICOLON MISSING IN x STORCB,w,u,D>. SHOULD BE x STOR(B,w,u;D
PUT, STORr OR MSAV TRIED TO WRITE TO PROTECTED AREA ON TAPE 10
DECTAPE/FLOPPY BLOCK NUMBER IS <O OR >1481
DECTAPE I/O AT'fEMPTED BEFORE FLOPPY I/O DONE OR ACKNOWLEDGED
ILLEGAL UNIT NUMBER SPECIFIED IN PUT, TAK, ASK, OR STOR
BI... 0 CI< NU MB E F< ~;; PE C I F I E D I N 1:· U T , T 1'~ !·\ , ,:~ E) I< , 0 F~ ~;; T CJ F\: I ~:; <0 0 i:;: >:l. 4 7 :
UNIT I CHANGED DURING USE OF PUT, TAK, ASK, OR STOR WITH B=W=O
WORD NUMBER < 0 SPECIFIED IN PUT, TAK, ASK, OR STOR.
WORD NUMBER > 190:1.45 SPECIFIED IN PUT, TAK, ASK, OR STOR
COMBINATION OF WORD AND BLOCK ts TOO BIG CPUT,TAK~AsK, OR STOR
NEGATIVE CHANNEL NUMBER SPECIFIED
NEGATIVE BUFFER NUMBER SPECIFIED
COMBINATION OF BUFFER & STARTING CHANNEL > AVAILABLE MEMORY
NUMBER OF CHANNELS TO PROCESS > TOTAL AVAILABLE CHANNELS
SAME AS EITHER 5103 OR 5105
ATTEMPT TO USE X NAME BUFFER 2 ON PDP8 WITH < 12K OF CORE
NEGATIVE OVERLAY NUMBER REQUESTED
LOCATION < 03200 (OCTAL) SPECIFIED IN X VAR
LOCATION > 04600 <OCTAL.> SPECIFIED IN X VAR
NEGATIVE PROGRAM i SPECIFIED IN X CALL., X FILE, DR X CADO
X CALL NESTING NO LONGER SUPPORTED. USE X CADO INSTEAD.
X CALL OR X END USED WITHIN A DO GROUP OR FOR LOOP
ATTEMPT TO USE X NAME BUFFER 2 WHEN THIS FEATURE IS DISABLED
:c1::iD C)',)EF<L_(l { F Dl.Ji'-1)) ClN T1:)p[ .... oi:;: Cl 1v'EF\'.1 ... {i y FOUND I'.:; NOT :·:)21< FUCt1!...
NON-EXISTENT PROGRAM CALLED WITH X CALL OR X CADO
TOO MANY DECTAPE/FLOPPY DISC ERRORS
AT'fEMPT TO USE FIELD 2 ROUTINES ON PDP8 WITH < 12K OF CORE

0455
2002
3336
4360
4411
4560
4567
4573
4615
4 ,;~ 2 3

4642
4647

'.5:1.03
'.'.'! 1 O'.~i
'.'5 :I. 10
'.'52 l (:)
c:• •""'t _. ..I 1:.~..t ..... ~ .~·:. \..l

:.s~:::(:;,3
~) .-:~;.

0 {1

54:~.'.4
:::;44;,;~

? () ~~~ ·.: -·.:? 0 ~:.=j ()
.• ., .-, 1::·
/

·~1

,.,·:. ~.) .,;)

.t" ... ) •••.::. ··:.•
.~·-

J /

THE FCJLLCJWING ERROR CODES PRINT OUT A ?2-NNNN?OO.OO
2-1006
2··.. I. 2<'.)0
2-2013
2-2120
2-2511
2-2542
2-5151.
2-7003
2-7011
2-7032
2-7363
2-7370
2-7432
2-7441
2-7444
2··.. 74~50
2-7453
2-7474

SINGLE PRECISION DIVISOR IN DMLJL IS <O OR >4095
1·1 I CPCJPHCJ TOMETEF;: Ot...'EF~L.1~~ Y FUNCTION u:~;ED Fl~:OM ~3C(iNNEF: FOC(1L
WORD COUNT > 2048 SPECIFIED IN X EXCR
DISASTER IN 32K TO 8K TAPE FORMAT TRANSLATCJR. REPOR'f THIS
ILLEGAL UNIT NUMBER SPECIFIED IN MGET OR MSAV
NON-EVEN WORD COUNT SPECIFIED IN X EXCR AND MODE NOT 0
GAP NOT FOUND WHEN WRITING IBM TAPE. CHECK FOR DIRTY CAPSTAN
NUMBER OF WOPDS <O IN X CID
t OF CHARACTERS )62 IN TYCO OR COTY, OR i OF WORDS >31 IN CID
i OF CHARACTERS <O IN TYCO OR COTY
NEGATIVE BUFFER NUMBER SPECIFIED
BUFFER SPECIFIED NOT AVAILABLE ON THIS SIZE MACHINE
ITAK OR IPUT TRIED TO READ OR WRITE 24-BIT t OFF ID BUFFER END
ID BUFFER WORD t < 0 IN IPUT, ITAK, TYCO, COTY, OR CID
ID BUFFER WORD t > 30 IN IPUT, ITAK• TYCO, COTY, OR CID
:C:UFFEF~ < .... :!. 3F'E:CIFIED IN IPUT, IT1:;1<, TYCOv CUTY:• Of< CID
BUFFER > 4 SPECIFIED IN IPUT, ITAK, TYCOv COTY, OR CID
LOCATION ( 0 OR > 17 SPECIFIED IN X PV OR GV

APPENDIX B3 - 32K FOCAL OVERLAY FUNCTION ERROR DIAGNOSTICS
THE FOLLOWING ERROR DIAGNOSTICS PRINT OUT EITHER AS
?NNNN?00.00 OR AS ?2-NNNN?OO.OO
CODE

MEANING

6045
6046
6051
6053
6055
6056
6060

AREA SCANNER NOT ATTACHED OR NOT TURNED ON
NUMBER OF CHANNELS < 0 OR > 4095 IN FIND
COEFFICIENT POSITION > 10 IN X SAV OR X COEF
SINGLE PRECISION DIVISOR IN x CORL rs < 0 OR > 4095
MINIMUM PEAK WIDTH < 0 OR > 4095 IN FIND
COEFFICIENT IN X SAV OR X COEF NOT PRECEDED BY
VALUE IS < 0 OR > 4095 IN X PUTW
X RDQ OR X WRQ WITH 'NO WAIT' OPTION FOUND IBM TAPE BUSY
STARTING WORD <W> < 0 IN X GETW, PUTW, COPY? RBLK OR WBLK
NEGATIVE DATA FOUND BY X LOOK
SMOOTHING FUNCTION OF LENGTH 0 SPECIFIED IN X CORL
PEAK CHANNEL > 4095 CHANNELS AWAY FROM STARTING CHANNEL IN FIN
INVALID UNIT IN RIT, LFTv UP1 DN
I OF CHANNELS TO PROCESS < SMOOTHING FUNCTION LENGTH IN X CORL
X POSF TRIED TO POSITION TO PROTECTED AREA OF FLOPPY ON UNIT 0
X BIN OUT OF ROOM FOR OUTPUT DATA
X POSF TRIED TD POSITION TO FLOPPY BLOCK <1
INTERNAL MICROPHOTOMETER SYSTEM ERROR. REPORT THIS.
GAP NOT FOUND IN X RJPL. CHECK FOR DIRTY IBM TAPE CAPSTAN.
ZERO CHANNELS TO PROCESS SPECIFIED IN X CORL
ORDER OF POLYNOMIAL > 9 SPECIFIED IN X POLN
CLOCK/CALENDAR SELECT CODE <O SPECIFIED IN FTIME
ORDER OF POLYNOMIAL > 10 SPECIFIED IN X POLY
CLOCK/CALENDAR SELECT CODE >5 SPECIFIED IN FTIME
INVALID AREA SCANNER ADDRESS IN X ASCR DR ASCW
AIR MASS*256 CZ> IS < 0 OR > 4095 IN X TINC
CHANNEL OFFSET (W) rs < 0 IN x SCRN
FIELD IS < 0 OR > 4095 IN FZAP, FJMP, OR FJMS
ANY OF THE ARGUMENTS IS < 0 DR > 4095 IN X SIZE
BUFFER 0 NOT ALLOWED IN X SET
PX, py, PPL, OR LPS IS =O IN X SIZE
I OF BITS TO SHIFT > 11 IN X COMB
X BIN RAN OUT OF UNBINNED INPUT DATA
A ZERO BIN SIZE IN BIN
PX*PPL*MU+XO IS > 1024 IN X SIZE
X SCRN RAN OUT OF UNSCRUNCHED DATA
STARTING WORD<S> <O OR >4095 IN X PAK OR X UNPK
STARTING WORD <W> < 0 IN X RJPL
NO WRITE RING IN IBM TAPE WHEN USING X WRQ
NEGATIVE BINSIZE FOUND BY X SCRN
BAD PIXEL SIZE IN X PIXL OR X SIZE
BAD COMBINATION OF Pl• PS, XO IN X PIXL
BAD COMBINATION OF LS• PS, YO IN X PIXL
X SET NOT CALLED BEFORE X CALB
VERIFY FAILED IN FZAP
INVALID MAILBOX MEMORY LOCATION SPECIFIED
INVALID FLOATING POINT BUFFER LOCATION
SEMICOLON MISSING IN x PUTVCS,B,v>. SHOULD BE x PLJTVcs,B;V>
I OF WORDS MODULOCI OF CHANNELS TO COMBINE> NOT 0 IN X COMB
INVALID OCTAL NUMBER IN FZAPv FDEC? FJMP~ FJMS1 DR FMPX
READ WILL SPILL OFF END OF MEMORY IN X RJPL
INVALID STARTING WORD <WA> SPECIFIED IN X ASTV
SCREEN FULL IN X IPHO OR X PIXL NOT CALLED FIRST

6062
6102
6105
6106
6113

6120
6124
6134
6137

6140
6153

6155
6160
6162
6165
6166
6170
6174
6204
6205

6206
6210
6216
6222
6231
6234
6244
6250
6251
6255
6270
6273
6274
6301

6302

6303
6311
6313
6317
6331

6332

6333
6335
6340
6343
6344

6362
6363
6364
6365
6366
6373
6412
6422

6434
6442
6445
6462
6472
6475
6520
6521
6537
6542
6545
6546

6547
6550
6556

6571

. ' _., 1
oou.

6603
6604

6607
6615
6626 .
6627
6644

6661
6665
6670
6703

6706
6711
6713
6714
6715
6725
6735
6742
6757

OF WORDS TO READ CNW> <= 0 IN X RJPL
ATTEMPT TO TAKE MICROPHOTOMETER DATA DIRECTLY TO DECTAPE
INVALID X COORDINATE IN X TVJY
INVALID Y COORDINATE IN X TVJY
INVALID MULTIPLIER FACTOR <MU> IN X TVJY
PY*LPS*MU+YO IS > 1024 IN X SIZE
DECTAPE I/O PENDING OR NOT ACKNOWLEDGED IN X RJPL
SCREEN FULL IN X PICT OR X SIZE NOT CALLED FIRST
INVALID STARTING BLOCK CSB> IN X RBLK OR WBLK
PATTERN BUFFER LENGTH CNSB> TOO BIG FOR PIXEL SIZE IN X PICT
ILLEGAL UNIT NUMBER SPECIFIED IN X RBLK OR WBLK
INVALID NUMBER OF BLOCKS CNB> IN X RBLK OR WBLK
DECIMAL NUMBER < 0 OR > 4095 IN FOCT
COMBINATION OF W AND NB SPILLS ACROSS FIELD IN X RBLK OR WBLK
X BIN OR INT CALLED WITHOUT CALLING X TABE FIRST
X HDTS OVERLAY REUSED~ OR QUICK IBM OVERLAY NOT LOADED
EXTERNAL GATE IN ONE STATE > 55 SECONDS IN X ASCT
INTERNAL MICROPHOTOMETER SYSTEM ERROR
INVALID STARTING ROW <SR) IN X ASTV OR TVOF
INVALID NUMBER OF ROWS CNR> IN X ASTV OR TVOF
INVALID NUMBER OF COLUMNS <NC) IN X ASTV OR TVOF
INVALID MULTIPLIER FACTOR IN X ASTV OR TVOF
I OF BITS TO SHIFT >= IN MAGNITUDE TO 24 IN X SHFT
FLOPPY I/O ERROR WHILE TAKING MICROPHOTOMETER DATA
AREA SCANNER MODE SET INCORRECTLY IN X ASCT
STARTING DATA WORD CWA> < 0 IN X HIST
DWELL TIME < 0 OR > 4095 IN X DTIM
DEADTIME FACTOR CD> IS < 0 OR > 4095 IN X DTIM
BUFFER 0 USED IN X SRCH
ILLEGAL MODE PARAMETER IN X ASCT
LOW OR HIGH BOUNDARY <LB OR HB) IS < 0 OR ? 4095 IN X HIST
HIGH BOUNDARY CHB) < LOW BOUNDARY CLB) IN X HIST
ODD NUMBER OF WORDS PER ROW CNC> IN X TVOF. MUST BE EVEN
NUMBER OF DATA WORDS CNW) < 0 IN X HIST
MICROPHOTOMETER DATA RATE EXCEEDS FLOPPY DATA RATE
PEAK WITH A WIDTH > 4095 CHANNELS FOUND BY FSEEK
SPILLED BACKWARDS OFF END OF MEMORY IN FLIP. REPORT THIS.
BIT NUMBER OF LENGTH <O OR >24 IN X BIT
INVALID UNIT IN X RDF OR WRF
END OF BUFFER SPACE IN X SRCH
X INT RAN OUT OF UNINTERPOLATED INPUT DATA
UNIT SPECIFIED IN X RDF OR WRF NOT POSITIONED WITH X POSF
i OF BLOCKS <O OR >31 SPECIFIED IN X RDF OR WRF
FSEEK RAN OUT OF OUTPUT BUFFER SPACE
X RDF OR WRF WITH 'NO WAif' OPTION FOUND FLOPPY BUSY
t

DECTAPE ERROR MESSAGE DECODING
MESSAGE

MEANING

T~PE?U---60--

MARK TRACK ERROR

TAPE?U---50--

END OF TAPE ERROR

TAPE?LJ---44--

SELECT ERROR

TAPE?U---42--

PARITY ERROR

TAPE?U---41--

TIMING ERROR

NOTES:
1. "U" INDICATES THE UNIT ON WHICH THE ERROR OCCURRED;
DECTAPE UNIT 8 IS INDICATED AS UNIT O. •-• INDICATES ANY
DIGIT BETWEEN 0 AND 7; THESE DIGITS ARE NOT SIGNIFICANT
IN DECODING THE MESSAGE.
2. SELECT ERRORS ARE THE MOST COMMON AND INDICATE A USER
ERROR. SELECT ERRORS ARE CAUSED IF THE DESIRED UNIT IS
NOT SELECTED~ OR IF ONE TRIES TO WRITE TO A UNIT WHICH
IS SWITCHED TO READ ONLY OR WRITE LOCK.
3. AN END OF TAPE ERROR IS ALSO PROBABLY A USER ERROR; IT MEANS
YOU TRIED TO ACCESS INFORMATION OFF THE END OF THE TAPE.
4. TIMING ERRORS INDICATE A HARDWARE PROBLEM AND SHOULD BE
REPORTED IMMEDIATELY.
5. MARK TRACK AND PARITY ERRORS MAY INDICATE A MARGINAL DISK;
IF THEY OCCUR REPEATEDLY, THE DISK OR TAPE IS PROBABLY
GOING BAD.

MULTIPLEXER ERROR MESSAGE DECODING

PRIVATE MULTIPLEXER UNIT NOT RESPONDING
PRIVATE MULTIPLEXER CHECK BITS TEST FAILED
PRIVATE MULTIPLEXER FUNCTION ECHO FAILED
PUBLIC MULTIPLEXER UNIT NOT RESPONDING
PUBLIC MULTIPLEXER CHECK BITS TEST FAILED
i'il.JX ff f' f 2 70 :!.

PUBLIC MULTIPLEXER FUNCTION ECHO FAILED
PUBLIC MULTIPLEXER BUSY BUS BUSY > 95 MILLISEC.
SAME AS MUXff ff5305

110 TE~:;;:

"ffff" IS THE OCTAL REPRESENTATION OF THE FUNCTION
WORD BEING TRANSMITTED WHEN fHE ERROR OCCURRED.
THIS FUNCTION WORD INDICATES WHICH MULTIPLEXER UNIT
WAS BEING ADDRESSED, AND WHAT OPERATION WAS BEING
PERFORMED. TO DECODE THE FUNCTION WORD, SEE PAGE 4
OF LICK OBSERVATORY TECHNICAL REPORT t:!.6,
"SERIAL DATA MULTIPLEXER - EL-331", BY T.P. RICKETTS
AND L.B. ROBINSON, FEBRUARY :!.976,

NOTE THAT MULTIPLEXER UNIT 3 CAN ONLY BE ADDRESSED
BY THE READ OUT ROOM COMPUTER, AND CANNOT BE
ADDRESSED BY THE ECHELLE COMPUTER. SIMILARLY,
MULTIPLEXER UNIT 4 CAN ONLY BE ADDRESSED BY THE
ECHELLE COMPUTER AND CANNOT BE ADDRESSED BY THE
READ OUT ROOM COMPUTER. MULTIPLEXER UNITS :!. AND
2 CAN BE ADDRESSED BY EITHER COMPUTER.

fURNING EITHER COMPUTER ON OR OFF CAN CAUSE THE
OTHER COMPUTER TD GET A FEW MULTIPLEXER ERRORS.
WHEN EI'fHER COMPUTER IS BEING USED FOR OBSERVING,
i:~ N Y

U N i'l E CE ~:; ~:; 1') P Y 1=· 0 kl E F'. I N C! U F' Cl r:~ D CJ l.<J N 0 F E I T H E F<

COMPUTER SHOULD BE AVOIDED.
4.

PLEASE REPORT ANY OF THE ABOVE LISTED MULTIPLEXER·
ERRORS TO BOB KIBRICK, TERRY RICKETTS, OR
LLOYD ROBINSON. PLEASE RECORD WHICH COMPUTER OR
COMPUTERS WERE IN OPERATION AT THE TIME THE ERROR
OCCURRED. ALSO RECORD WHICH DATA TAKING SYSTEM OR
SYSTEMS WERE IN USE.