MEMOTECH MEMOCALC Spreadsheet Analysis

Hello User!

More and more people are using the ZXE1 in business and engineering. Systems of all kinds require planning, and planning recuires a model of the future (or possible futures). This is where MEMOCALC comes in. You can now use powerful and rapid projection techniques on your ZX81

In this booklet you'll find all the functions and error messages explained, and some worked examples to demonstrate these.

Good luck from all at Memotech!

What is spreadsheet analysis?

Spreadsheet analysis is a system for entering data and equations relating to those data in a common rectangular grid. When the grid is eva uated, each equation is solved to give a value which then appears in the cell where the equation was entered.

What then is MEMOCALC?

MEMOCALC is a spreadsheet analysis package with a grid size of up to 7000 numbers. It lets you use all the functions of the ZX81 to build up simple or sophisticated models for business, engineeriny and other applications.

The program comes in the form of an EPROM pack that fits snugly on the back of the ZX81. No time is wasted loading the program, and all your BASIC addressable memory is available for holding grid data.

The program is easy to use once the basic principles are grasped. All inputs are prompted and if an error occurs a meaningful message is given. What's more, column and line headings are shown on the screen, so you always know where you are.

You can display figures with two decimal places or with none (but the values are held internally with greater accuracy to prevent the build-up of rounding errors). There are totalling and sub-totalling facilities, and a special feature to enable

the months of the year to be put in rapidly.

You may want to print the results of your model. As well as supporting the ZX printer, MEMOCALC can be used with other printers through the MEMOTECH printer interfaces to give your results a truly professional appearance.

How do I get started?

First set up your configuration with at least the ZX81, the MEMOCALC pack and a MEMOPAK or Sinclair 16K RAM pack in that order. MEMOCALC is addressed between 12 and 16K. You can either switch on your MEMOCALC by hand (at the side of the pack) to jump straight in, or make a USR call at 13824 (eg RAND USR 138324) after you have powered up. Note that if you have a Memopak HRG in your configuration with Memocalc, the HRG will only be accessible when Memocalc is switched off.

Immediately you will be prompted to choose between L for LOAD (to read an existing grid from tape) or I for INIT (to start afresh). For your first time around you will need to key in 1, whereupon the prompt NO. OF LINES? will appear.

What does the spreadsheet grid look like?

Unless your grid is very smalJl, you will never see it all at once, but the screen acts like a window through which you can see 3 columns and 15 lines at a time. At the top and down the left hand side are the text areas, separated from the main body of the grid by broad black lines. The cursor is a bar of 7 black squares, and by pressing 5,6,7 or 8 you can move it around the grid in the direction of the associated arrow. If you 'shift' these keys, the window moves around the grid and

the cursor rnoves with the window.

What can I do with tho grid?

There are a number of functions you can call by entering the first letter. If you enter H for the HELP function you will see what they are. The letter will be stored in the top lefthand corner of the screen to remind you which function you are using. One of the main function codes is I for INPUT. If you press I when the cursor is in the text area, you will be prompted TEXT?; you can then enter a label for a line or column. If your cursor is situated in the main body of the grid then you will be prompted with VALUE/EQUATION? V is for VALUEs or raw numbers, E is for EQUATIONs for generating new values from previous ones. There is also R for REPEAT and T for TOTAL which are discussed below, and C for CALCULATE. If you get into a tangle, shifted 1 (EDIT) will abandon the current function.

How do I enter positions and equations?

Positions are defined by grid references with numbers representing the columns, and letters representing the lines. For example, A1 is Line A Col 1; C8 is Line C Col 8. Note that when all the single letters have been used up then double letters are used, i.e. AA is Line 27, AK Line 28 and IO is Line 250 - the last possible line. Whenever the program prompts POSITION? you shouid supply the grid reference in this form.

Equations are expressed using position references as variables, the left hand side of the expression being the cursor location. For example, if you want to assign to cell A1 twice the value of cell D12, you should move the cursor to A1 (using G (GOTO) or the 5, 6, 7, 8 keys) and then key in I (INPUT), E (EQUATION), and D12 * 2.

Equations can be self referential. To set a initial value to a cell, use the I V function; an equation can then be entered in the same cell referring to itself. It will immediately evaluate itself. For an example of this, see CALCULATE in the List of Functions below.

A MINI-EXAMPLE

Green the Grocer has a value for each kind of produce he thinks he will sell in January. He thinks apples will then increase each month by 10%, pears will decrease by 5%, sprouts will double and leeks halve. He would like to examine

the overall situation for the quarter, with separate monthJly sub totals made on fruit and vegetables, and totats and running totals for each month.

After I for INIT he enters 4 for NO. OF COLUMNS? (1 per month and 1 for a quarter's totals) and 15 (the minimum) for NO. OF LINES? He then proceeds to set up his grid labels. In each case he first moves the cursor to the appropriate position, then enters his function(s) and then the text, followed by NEWLINE. There are two rows available for labelling columns (called here 'upper' and 'lower' ) .

CURSOR POSITION FUNCTION(S) ENTRY

TYPED IN

Upper label col 1 | GREEN

Upper label col 2 | THE

Upper label col 3 | GROCER

Lower label col 1 | JAN

Lower label col 2 | FEB

Lower label col 3 | MAR

Label line A | APPLES

Label line B | PEARS

Label line C(left blank) |

Label line D | FRUIT

Label line E(left blank) |

Label line F | SPROUTS

Label line G | LEEKS

Label line H(left blank) |

Label line I | VEG

Label line J(left blank) |

Label line K | TOTAL

Label line L(left blank) |

Label line M | RUNNING

That completes the labelling.

Say his initial values for apples, pears, sprouts and leeks are 100, 150, 200 and 250 respectively. He proceeds:

CURSOR POSITION

A1 I V 100

B1 I V 150

F1 I V 200

G1 I V 250

100 150 200 250

Next he enters equations by which the program will generate the future values,

A2 I E A1*1.1

B2 I E B1*.95

F2 I E F1*2

G2 I E G1*.5

A2 R A3

B2 R B3

F2 R F3

G2 R G3

Note that Mr Green has saved time by using the R (REPEAT) function: with the cursor at A2, R A3 means repeat the equation which is in A2 along line A as far as A3. After the C O (CALCULATE ONCE) function has been performed, a D (DECODE) function with the cursor at A3 will reveal A2*1.1

He wants line totals generated in column 4, so he first moves the window:

Shifted 8 (moves window to the right by 1 column)

Lower label col 4 | QTR

A4 T T L

B4 T T L

F4 T T L

G4 T T L

Shifted 5 (moves window to the left by 1 column)

And lastly the column sub totals and totals.

D1 T S C

D1 R D4

I1 T S C

I1 R I4

K1 T T C

K1 R K4

M1 I E K1

M2 I E M1 + K2

M2 R M3

At the end, or at any point in the input procedure, he can enter C for

CALCULATE to see the results, followed by the letter O since he only wants

to calculate ONCE. After any amendments, C should be re-entered to see the

new values.

The final result should look something like this:

memocalc GREEN THE GROCER

JAN FEB MAR QTR

1 2 3 4

APPLES a 100 110 121 331

PEARS b 150 142 135 427

c

FRUIT d 250 252 256 758

e

SPROUTS f 200 400 800 1400

LEEKS g 250 125 62 437

h

VEG i 450 525 862 1837

i

TOTAL k 700 777 1118 2596

I

RUNNING m 700 1477 2596

n

o

What form do the entries take?

All inputs are in response to prompts and can be of two types - single characters or strings. If a single character is required no cursor will appear at the bottom of the screen, and afterwards there's no need to press NEWLINE. Sometimes you are asked to make a choice of two letters shown on a black background Just press the key corresponding to the desired letter.

If a string is required, an L cursor will appear at the base of the screen. As you enter the string it is displayed below. Afterwards you must press NEWLINE before the computer wil pay any attention to it. You must enter at least one character before you press NEWLINE. The program may limit the characters you are permitted (eg for numbers it only allows keys 0-9). Normal ZX81 editing features are supported.

At the base of the screen, input can be up to 4 lines long. If you go over 4 lines, the input will be rejected and TOO LONG displayed. If the input has format or syntax errors, it will be redisplayed with a marker at the error which can then be edited or re-entered. If an input is syntactically correct but cannot logically be implemented in the grids it will be rejected with an error message.

LIST OF FUNCTIONS

SHIFTED 1 (EDIT)-abandons current function .

UNSIFIFTED 5, 6, 7 AND 8 moves the cursor around the screen, in the direction shown on the arrow.

SHIFTED 5, 6, 7 AND 8- moves the 'window' on the data that is currently being displayed .

B - BLANK blanks out the value/equation/subtotal/total at the position on the screen occupied by the cursor. For the purposes of calculating, blank positions have a value of zero.

C-CALCULATE forces a recalculation of all values in the matrix (ie evaluates all the equations, subtotals and totals) A prompt of ONCE/MANY? will be displayed Press O if you want the matrix to be evaluted just once. Note that calculations are done from left to right within an overall top-to-bottom progression. You can only evaluate equations from entered values or from results of previous equations. You should ensure that equations only refer to the positions above or on the same line to the left, unless the values are fixed. A line of running totals may easily be accumulated beneath a line of absolute column totals, but carrying forward from the base of one column to the top of another is not possible with the O option.

Pressing M for MANY causes the matrix to be evaluated and the bottom right cell to be tested for its value. If this value is exactly 0 (and a blank will count as 0), the cycle is deemed to be finished and the calculation will stop. Entering a non-zero VALUE in that cell will cause the matrix to be re-evaluated cyclically until the BREAK key is pressed. Using that cell only, a simple counter can be inserted to give a known number of cycles; for example, in a 3-column; 15-line matrix the entries at cell 03

I V 21

I E03-1

will cause the matrix to be evaluated 20 times and then stop.

D-DECODE tells you how the value at the cursor position is derived. This will either be an equation or the codes BL-blank/ VL-value, SC-subtotal column, SL-subtotal line, TC total column, or TL-total line.

G-GOTO moves the 'window' on the grid so that the top left hand corner is the line/column you supply in the response to the POSITION? prompt. The cursor will be over the position supplied.

H-HELP lists available functions; these appear beneath the grid and are removed when another function is selected.

I- INPUT enters information where the cursor is resting. If this is in a heading position (above or to the left of the grid) then text will be expected and the prompt TEXT? wiil be displayed. In text mode all characters (but not graphics symbols} are acceptabie; only the first 7 are used.

If the cursor is resting over the body of the table, then the prompt EQUATION/VALUE? will be displayed. If the entry in the cell is not to be dependent on other values in the table, then the VALUE option should be used. Otherwise use EQUATION. In response to these choices the prompts VALUE? or EQUATION? respectively will be displayed.

In the case of EQUATION a valid mathematical expression should be entered using position references as variables. The RND function is supported as are the logical operators (AND, OR and NOT). Upon entry, the program will check the equation for vaXidity and redisplay any invalid expression with an error marker and with the cursor positioned after the offending character. If the expression is valid, then its results will be calculated and displayed immediately and every time the CALCULATE command is given. Note that EQUATIONs can be copied into other positions in the grid, whereac VALUEs cannot. To effectively copy a vaiue, enter it first as an EQUATION but without position reference variables.

L- LOAD loads a grid from tape or sets up a new grid. Any grid already in RAM is lost. The prompt LOAD/INIT? will be displayed.

If you press I, existing data will be erased immediately. The prompt NO. OF COLUMNS? will then appear. Reply with the number of columns of data you expect to use (leave some room for expansion - big empty matrices take longer to SAVE than small ones, but no longer to CALCULATE). The number of columns must be in range 3-99. Following this, the prompt NO. OF LINES? will appear. The permissible range is 15-250.

The program will then see whether there is space in memory for an array of this size. Generally, lines x columns should not exceed 1800 for 16K, 4400 for 32K, and 7000 for the 48K RAM in the Memopak 64K. If your demands exceed the space available, you will be asked for NO. OF COLUMNS? again.

If you choose the L (LOAD) option, you will be prompted for the NAME? of the grid you wish to load; this must be exactly the same as the name it was saved under. There is no default loading of the first thing on the tape, so remember to record the names you assign to your grids. Start the cassette recorder and press NEWLINE; when you press this key, any existing data in RAM will be cleared The screen will go blank for a few seconds when the load is completed while the display is built up. If an error occurs in the load you will be returned to the LOAD/INIT? prompt so that you can try again.

M - MODE switches between integer and 2-decimal display modes. Integer display occurs initially, but you can switch between modes at any time. Existing data on the screen is not immediately changed, but new entries are in the new mode. When the screen is re-displayed (on CALCULATE or GOTO) the new mode is employed.

Note that the largest number that can be displayed is 9999999 in integer mode and 9999.99 in 2-decimal mode. Although larger numbers can be handled and stored, they will overwrite the previous column in the display and printout. If you think this will be a problem, just use every other column in the grid, though you will then be restricted to column REPEATs only.

P- PRINT elicits the SCREEN/ALL? prompt. Pressing S will cause what is on the screen to be copied to the printer. A for ALL will ask you for the NO OF COLUMNS? you want printed across on your printout; the maximum number of columns you can print across a sheet is as follows:

ZX printer 3 columns

80 character printer 9 columns

132 character printer 15 columns.

Either print option can be halted by pressing the BREAK key.

Q- QUICK Converting between internal floating point numbers and displayed decimals is a slow process. In QUICK mode, existing data are not displayed although labels are, and shifting the window becomes viritually instantaneous. All data are however displayed after CALCULATE or GOTO. To get back out of OUICK mode, simply press C again .

R - REPEAT repeats an equation (or total/ subtotal) from the cell where the cursor is resting to another position which must be either on the same line or in the same column. Note that the repetition is relative to position, i e if the initial equation is A1*2 and you are REPEATing it downwards, the DECODE function would reveal B1*2 in the next cell, then C1*2, and so on.

S - SAVE saves the entries in the grid. In response to NAME? you should enter a name of up to 12 characters. You do not need quote marks as in the normal ZX81 SAVE. Escape from the SAVE can be effected using the BREAK key

T-TOTAL/SUBTOTAL works down a column or across a line. You will be prompted TOTAL/SUB-TOTAL? and should respond T or S accordingly. Then the LINE/COLUMN? prompt will appear; responding L accumulates values to the left on your line, and C accumulates above in your column. For examples T S C means TOTAL function/ Subtotal of Column. It works upwards accumulating values in the column until it reaches either a previous column subtotal or the first line. It ignores any column totals but will include line subtotals and totals.

Y - YEAR provides a fast method of entering the months of the year as labels. The cusorshould be positioned over a label cell containing the first three letters of a month (JAN,FEB, etc.). When Y is pressed the remaining months of the year will be set up in the succeeding cells (to the right if a column label, downwards if a line label). If the initial cell contains some other information in the four remaining character spaces (eg JAN'84), that information will be copied with the succeeding months. The effects of the Y function are immediate, but are not displayed until a CALCULATE or GOTO command is given.

ERROR MESSAGES

CALC. ERROR (functions C, I & R)-A mathematical error was detected during processing. The line/column reference at which the error was detected is given, and this is followed by the ZX8l error code (see Appendix B of the Sinclair manual). If function R is being used then the chances are the equation you are trying to repeat uses grid references and that there are no corresponding grid references in the repeat-to position. ( Function D can often be used profitably when this occurs to help you track down the source of error).

CANT EXTEND (function Y) The box over which the cursor is resting does not contain a recognised month of the year (or contains DEC which cannot be extended), or the dimensions of the grid do not allow any extension to take place.

EDIT USED (any function) - The EDIT key was pressed during INPUT. Nothing you have done so far within this function has affected the data.

SAVE ERROR (function S) - An error occurred on the SAVE (or you pressed the BREAK key to terminate it intentionally).

INV.REPEAT POS ( function R ) - The REPEAT position is invalid, either because both line and column references of this posistion are different to the cursor location, or because this position is not below or to the right of the cursor location.

NOT A NO (functions B, D, R, T) -The cursor is resting over a header field, so that these functions which are concerned with the body of the grid cannot be used. (To blank out a value in a header field use function I and key in spaces).

NOT EQUATION (function R) The cursor is resting over a position which contains a VALUE not an EQUATION. Such values cannot be repeated. (If you want to REPEAT a value you should set it up as an equation without position references).

NOT TEXT (function Y)-The cursor is not resting over a header field.

OUT OF RANGE (functions L, P and during the initial entry to the program) If you are using function P, then the number you entered after the NO. OF COLUMNS? prompt was not in the range 1-15. When setting up the dimensions of the model, the number of columns must be in the range 3-99, and the number of lines in the range 15-250.

TOO BIG-REINPUT - There is not enough room in memory to accommodate the number of columns and lines you have requested (and still leave some room for equations) .

TOO LONG (function I) - This error message arises a) because the attempted INPUT has exceeded 4 lines in length, or b) because no more room exists for equations. Room can be cleared by BLANKing out unwanted entries.

A BIGGER EXAMPLE

You are a manufacturer who expects sales of 1000 units of value in January, rising by 1% per month for the year. Your nett costs are 90% of sales value. You want to know your profit and accumulated profit at each month.

Press I for INIT, 4 for NO. OF COLUMNS? (to allow for sales, costs, profit, and accumulated profit) and NEWLINE, and 15 for NO. OF LINES? (which is the minimum, and allows for 12 months and a total) and NEWLINE.

The empty grid will come up, showing columns 1-3 and lines A-O; the cursor will be at A1, and the prompt FUNCTION? will be at the base. Remember that if you want to cancel the function you're in for whatever reason, you can get out with EDIT (shifted 1 ).

Before you enter the function command move the cursor.

Put the cursor above the black line which marks the edge of the table so that when you enter I the prompt TEXT? appears. Enter SALES, and move the cursor across for COSTS in column 2, and PROFIT in column 3. In order to enter CUM in column 4 you will have to shift the 'window' using shifted 8. When this is done you can GOTO A1 again.

Move the cursor to the left to enter the label for line A: enter JAN 84. You could then enter all the other months of the year yourself, but to make it easier leave the cursor where it is and enter function Y. Now C (CALCULATE) and see the effect of that.

The cursor will now be back at A1 and you can enter the initial value of sales. Press I V 1000 and NEWLINE.

Sales for FEB are a 1% increase over JAN's, so move the cursor to A2 and enter I E 1.01-A1. When you press NEWLINE the value 1010 will appear there.

You could now enter a similar equation at C1, which would be 1.01+B1, and so on for all the other months. The R (REPEAT) function is designed to save you this trouble; it gives you the prompt POSITION? which asks you how far down the column you would like the pattern of the equation to be repeated. This pattern is relatively positioned, so the equation is understood to mean 'multiply the value above by 1.01 and place the result here'. Leave the cursor at B1 and enter R. In response to the prompt, enter L1 (corresponding to DEC 84), and CALCULATE, ONCE.

If you're interested, move the cursor to I1 and DECODE. Press M for MODE and CALCULATE. Press H and see what else there is.

To get on with the COSTS column, move the cursor to A2. Costs are 90% of sales value, so enter I E 0.9*A1, R to L2, and C O.

Profits are sales minus costs, so move the cursor to A3. The entries here are I E A1-A2, R L3, C O.

To see column 4, use shifted 8. JAN's cumulative profit is simply the profit in JAN, so with the cursor at A4 enter I E A3. For the second and subsequent months, the cumulative profit is the previous cumulative profit plus current profit. Move the cursor down to B4 and enter I E A4+B3, and REPEAT down to L4. C O will give you the results as usual.

All that remains now is to produce annual totals for the columns: these can go in line N. Entering G (GOTO) N1 will put the cursor where you want it, and T T C will give yot your column total. Until you CALCULATE, 0 will appear there to show that something has been accepted. R N3 and C O will complete this spreadsheet. You should get a total profit for the year of 1268.

Sometimes figures do not appear to add up. This will be because values are always rounded down for display. Function M will change from integer to decimal (or decimal to integer) modes. Then you have to key G for GOTO and reply A1 to POSITION? to see the more accurate figures. (G for COTO is very useful to move among larger models).

If this were a real model, you might want to SAVE it or PRINT it. See the list of functions. If you like you can change the value in A1 and re-calculate the whole grid. Or you could try and cycle the program to show the profit after 5 years.

STRANGE RESULTS

If you get results that are unexpected, then almost certainly something is wrong with the equations. You should check that equations refer only to grid references where the value at that cell has been evaluated first (remember that the calculation works from left to right and from top to bottom), and that are not set to blank.

HINTS ON MATHEMATICAL FUNCTIONS

MEMOCALC supports all the mathematical functions of the ZX81. This makes it ideal fo engineering applications as well as financial projections.

The RND function

This is an important simulation device. Looking at the last worked example, we assumed that sales increased smoothly. This is ver unlikely to happen - there may be a underlying upward trend, but sales in each month may show peaks and troughs.

If sales are expected to be random within plus or minus 10% of the expected sales, we could model this quite simply with an equation such as:

Random sales= Expected sales * (.9+ .2 * R N D)

More sophisticated equations can be used to model statistical distributions, or to reflect relationships between monthly sales. (If, for example, you had a good sales month in January, it may well be that some of February's sales came early).

The advantage of incorporating RND in such equations is that you will get different random values every time you calculate, and apart from pressing C you do not have to put in much effort. Done several times, this may reveal potential problem outcomes that you might not have considered.

LOGICAL OPERATORS

AND, OR and NOT can be used to put conditions into the model. The costs in the worked example are to say the least, simplistic. A more realistic situation might be that there are fixed non-production costs of 300, a material cost of 0.3 per unit sales, and that factory; labour and overhead costs are a minimum of 300 no matter how many are being made; but that anything over 1000 units produced will incur an excess factory cost of 0.4 per unit. We can then write a mathematical expression for the costs as:

(300+ 0.3 * SALES + 300) + 0.4 * (SALES-1000) AND (SALES >1000)

The right hand part of this expression means that a cost of 0.4 per unit is incurred on sales in excess of 1000 only if it is true that saies exceed 1000.

Such instructions are very useful in incorporating differential tax bands or interest rates etc. Chapter 10 of the Sinclair ZX81 manual goes into more detail.

MEMOCALC was written by John Reidy for MEMOTECH LIMITED.

MEMOCALC (c) R. BRANTON & G. A. C. BOYD 1982