1 Start Here

If you bought the ATOM ready built, together with a power supply and a cable to connect it to a TV set, then carry on reading. Otherwise you should refer to the Technical Manual for details of how to assemble an ATOM kit, and for details of the required accessories. The ATOM connects to the aerial socket of an ordinary black-and-white or colour TV set. The ATOM will not affect the normal operation of the TV in any way. Connect the UHF output from the ATOM to the aerial socket of the TV set; see Fig. 1. Connect the ATOM's power supply to a mains socket, and plug the power connector into the back of the ATOM; again, see Fig. 1. Press the key marked BREAK on the top right of the ATOM's keyboard. Switch on the TV set, and turn the set's volume control down. The ATOM makes use af a TV channel that is not occupied by any TV stations, and it is necessary to tune to this channel in order to get the correct display from the ATOM. If the TV set you are usinq has push buttons to select stations, choose an unused button and tune the TV by rotating the button. If the TV has a single numbered tuning dial, turn the dial to somewhere near channel 36. Tune in the TV set until the screen is black, with the following display in the top left-hand corner of the screen: ACORN ATOM <_ Adjust the contrast and brightness controls so that the letters are clearly legible, and tune the TV set carefully until the letters are sharp and clear. The '<' sign is called the ATOM's 'prompt'. It indicates that the computer is waiting for something to be typed in; a command, perhaps, or a program. The white rectangle, '_', is called the 'cursor'; it indicates where on the screen the next character you type in will appear.


l.l What the ATOM Can Do The ATOM understands the following special words and symbols: Commands LIST, LOAD, NEW. Functions ABS, BGET, CH, COUNT, EXT, FIN, FOUT, GET, LEN, PTR, RND, TOP. Connectives AND, OR, STEP, THEN, TO. _3_
Statements BPUT, CLEAR, DIM, DO, DRAW, END, FOR, GOSUB, GOTO, IF, INPUT, LET, LINK, MOVE, NEXT, OLD, PLOT, PRINT, PUT, REM, RETURN, RUN, SAVE, SGET, SHUT, SPUT, UNTIL, WAIT. Operators !, #, $, &, *, +, -, /, :, < =, >, ?, \, <>, < >=. These words and symbols will be explained over the course of the next 12 chapters; for the moment just observe that many of these words have an obvious meaning; for example, try typing: PRINT "HELLO" after the '>' prompt sign. Note that the quotation marks are obtained by holding down the SHIFT key and typing the '2' key. Now type RETURN to indicate that the line is finished, and the ATOM will do just that: HELLO>_ To perform calculations you just need to type PRINT followed by the expression you want to evaluate. For example, try: PRINT 7+6*2 When you type RETURN the answer will be printed out. You can try typing anything you like, but any words not on the above lists will probably cause an error. For example, try typinq: HELLO after the ATOM's ','’ prompt. The ATOM will reply with a 'bleep' and will print: ERROR 94 which means that HELLO is not one of the statements or commands that the ATOM understands.


1.2 A Demonstration Now that you are in control of your ATOM you may like a quick demonstration of some more complicated things that it can do. No attempt is made here to explain how these examples work; for that you will have to read the rest of the first section of this manual. You can make ATOM do a lot of typing with very little effort; try entering: DO PRINT "ATOM-"; UNTIL 0 Note the difference between the '0' of DO, which is the letter '0', and the '0' at the end of the statement, which is the digit '0' on the top row of the keyboard. You will have to type the ESC (escape) key, which is at the top left of the keyboard, to stop this program. Now try typing in the following line: DO PRINT $RND&3+8,$8,$128; UNTIL 0 You will need to use the SHIFT key to get some of the special symbols. This program is longer than one line of the screen, but just keep typing and it will appear on a second line. Then press RETURN to run the program. Again, you will have to type ESC to stop this program. To demonstrate the graphics commands type: CLEARO; MOVE 10,0; DRAW 60,50 and the ATOM will draw a line on the screen. If you feel like trying a more complicated graphics program, type in the following: _4_
CLEAR0;MOVE32,24;Y=l;DOPLOT1,0,Y;PLOT1,Y,O;Y=-Y-2*Y/A.Y;U.O Press ESC to get back the ATOM's prompt. To demonstrate the ATOM's assembler enter the following line after the prompt: P=320;[INX; LDA 0,X; STA 4B002; JMP 320;] An assembler listing will be printed out, and the machine code will be put into memory at 320. To execute the program, type: LINK 320 and the ATOM will make a buzzing noise. It is playing the random contents of its memory through its internal loudspeaker. To stop the program you will have to type BREAK, because it is a machine-code program. You may question the usefulness of these examples, but they do illustrate the wide range of different tasks the ATOM is capable of. These 'programs' all fitted onto two lines of the display; to see what you will be able to do with a longer program take a look at the many examples later on in this manual.


1.3 The Keyboard The ATOM keyboard is designed to the standard layout generally accepted in the computer industry; see Fig. 2. In most respects it is just like the keyboard of an ordinary typewriter, but there are some important differences. For a start there are several keys not found on typewriters, such as DELETE, REPT, CTRL, and BREAK. The purpose of each of these keys will be explained in the following sections. Another difference is that the letters A-Z will appear in capitals, rather than lower-case, when they are typed by themselves. Try typing in the letters 'ABC' and observe that they appear, as you type them, on the screen: >ABC_ From now on, in the examples, the cursor will not be drawn in for simplicity.


1.3.1 SHIFT Some keys carry two legends. For example, each digit key (except 0) also has a special symbol or punctuation mark above it. The lower symbol on each of these keys is obtained by simply typing that key; the upper symbol is obtained by holding one of the SHIFT keys down, and typing that key. This aspect of the keyboard is just like a typewriter. If the SHIFT key is held down in conjunction with one of the keys bearing a single legend, such as A-Z and 0, [, etc, then the character will appear inverted; i.e. as a black character on a white square. Inverted A-Z correspond to lower case letters, and will be represented by lower case letters a-z in this manual. Inverted 0, [, \ etc. will be represented by 0, g, g etc.


1.3.2 LOCK The LOCK key, when pressed on its own, changes the way the SHIFT key operates with the letters A-Z. Initially the keyboard will give inverted A-Z in conjunction with the SHIFT key, and plain A-Z otherwise. If the LOCK key is now pressed once the keyboard will normally give inverted A-Z, and will give plain A-Z only when the SHIFT key is held down. Pressing LOCK again will revert to the previous state. _5_


1.3.3 DELETE The advantage of a TV screen over a piece of paper is that mistakes can be corrected without trace of the error. The DELETE key will erase the last character on the line, and the cursor will back up one space. Mistakes can thus be deleted and retyped with little effort.


1.3.4 RETURN The RETURN key is a signal to the computer that you have finished typing in a line of characters. The cursor will move to the start of the next line, and the computer may respond to what you have typed by typing out a reply.


1.3.5 Repeat - REPT If the 'repeat' key, marked REPT, is held down with another key, that key is typed repeatedly. REPT is useful in conjunction with DELETE to erase several characters very rapidly. Note that pressing REPT on its own will have no effect.


1.3.6 Control - CTRL There are several special functions available from the keyboard which are obtained by typing certain keys with the 'control' key - marked CTRL - held down. Only the following two control functions will be mentioned here: CTRL-G gives a bleep in the ATOM's loudspeaker. CTRL-L clears the screen.


1.3.7 BREAK The BREAK key will reset the computer, and return it to the state it was in just after switching on. It should not normally be necessary to type BREAK, but some assembler programs can cause loops which cannot be stopped in any other way. Note that the contents of memory are preserved when BREAK is typed, and any stored program can be recovered.


1.4 Scrolling When the cursor reaches the bottom of the screen further lines typed in will cause the screen to 'scroll'; every line is shifted up so that you always see the last 16 lines of what has been typed, and the top line of text on the screen will be lost.


1.5 Storing Text Any line typed after the ATOM's '>' prompt which starts with a number is not executed, but stored as text in the ATOM's memory. Any type of input can be stored in this way; it could be the text of a document, a program in BASIC, an assembler program, or data for a program. This section shows how to enter a piece of text, which can then be stored on cassette, edited, or output to a printer. The same method would be used for entering a program. The line must start with a line number, which can be any number within the range 1 to 32767, and there is no need to use consecutive line numbers for consecutive lines; indeed, it is wise to choose line numbers spaced by about 10 as you will soon realise. After the line number you should type the line of text. For example, enter the following: _6_
10 IN XANADU DID KUBLA KHAN 20 A STATELY PLEASURE-DOME DECREE: 30 WHERE ALPH, THE SACRED RIVER, RAN 40 DOWN TO A SUNLESS SEA. Remember to type RETURN at the end of each line. Each line number can be followed by up to 64 characters; if you try to type more than 64 characters the ATOM will refuse to proceed until you have deleted some characters. The reason for spacing line numbers somewhat apart is that it is then a simple matter to insert new lines between existing lines. For example, to insert a line before line 40, type: 36 THROUGH CAVERNS MEASURELESS TO MAN The computer will sort the lines into the right order, according to their line numbers, irrespective of the order in which you entered them.


1.6 Commands Commands typed in after the '>' prompt, without a preceding line number, and followed by RETURN, are executed immediately by ATOM rather than being stored in its memory. For example, now type the command: LIST This will cause the stored text to be typed out: 10 IN XANADU DID KUBLA KHAN 20 A STATELY PLEASURE-DOME DECREE: 30 WHERE ALPH, THE SACRED RIVER, RAN 36 THROUGH CAVERNS MEASURELESS TO MAN 40 DOWN TO A SUNLESS SEA. There are several options with the LIST command. For example: LIST 10 will list line 10 only. LIST 20,40 will list lines 20 to 40 inclusive. LIST 20, will list line 20 onwards. LIST ,30 will list up to line 30. A listing can be stopped by typing ESC (escape).


1.7 Editing One powerful feature of the ATOM's text and program storage is that stored lines can be modified very simply by typing the same line number followed by the new version. For example, to change line 20 in the text just type: 20 NEW LINE TWO and try listing the program again to see the effect. To delete a line simply type the line number followed by RETURN.


1.8 Other Commands Some other useful commands are described here: NEW will clear the stored text so that a new piece of text can be typed in. It should always be typed before entering a new piece of text. OLD can be typed after typing BREAK to retrieve the text previously in _7_
memory. Note that you should only type OLD if there is already text in memory.


1.9 Errors By now you the ATOM will probably have made a 'bleep' followed by the message: ERROR X where X is the error code number. There are two possible reasons for errors: 1. You typed something, probably a command, that the ATOM was not expecting or could not interpret. 2. The ATOM was commanded to do something that it could not do. For example, typing 'ABC' followed by a RETURN will give the error message: ERROR 94 which is probably the most common error; it means that 'ABC' was not a legal command. Remember that it is impossible to cause physical damage to the ATOM, whatever you type at the keyboard. The worst you can do is to lose the stored text, and even that is extremely unlikely. Most errors are really warnings, and a complete explanation of all the error codes is given in Chapter 27.


1.10 Saving Text or Programs on Tape Having entered some stored text into the ATOM's memory, this section will show how to save this text, and load it back at a later time. Text and programs can be saved on standard cassette (or reel-to-reel) tapes using the ATOM's cassette interface. Connect the cassette output from the ATOM to the input of a cassette recorder, and the output from tne recorder to the input of the ATOM. The tape load routine uses software averaging techniques to minimise the likelihood of errors on loading, and no trouble should be experienced in transferring tapes from one machine to another.


1.10.1 Setting Up Before loading and saving files using the cassette interface it is worth entering the following simple routines to check that the cassette system is working correctly, and to find out the best setting of the recorder's volume control. Enter the following line after the ATON's prompt: DO BPUT A,88; WAIT; WAIT; WAIT; WAIT; UNTIL 0 Type RETURN and record on the recorder for a few minutes. To stop the program type ESC (escape). This program has recorded a sequence of Xs, in coded form, on the tape. If you play it back it should sound like a series of short buzzes. Now enter the following line, which is a program to read characters from the tape and print them on the screen: DO PRINT Sbget A; UNTIL 0 The dollar symbol is obtained by holding the SHIFT key down and typing '4'. Press RETURN, rewind the tape, and play back the 'X's that you recorded. If all is well a stream of 'X's should be printed out, and adjust the volume setting on the recorder so that no other characters appear, indicating errors. When you are satisfied that all is well, _8_
proceed to the next section.


1.10.2 Text Files The information is stored as a stream of audible tones on tape; each section of information is referred to as a 'file'. Several different files can be saved on one tape, and they are identified by having unique 'filenames'. Filenames can be anything containing up to 16 letters, digits, or spaces: suitable names are "DATA FILE", "22/4/80", etc.


1.10.3 SAVE First check that the stored text is still there by typing LIST. To save the stored text to tape, type: SAVE "EXAMPLE" where "EXAMPLE" is the filename chosen for illustration. Type RETURN, and the message: RECORD TAPE will be printed on the screen. Put the tape recorder to record, and allow the tape to run well past the leader. Now type RETURN (or any other key) and the cursor will move to the start of the next line, indicating that the text is being recorded. After a short delay the '>' prompt will reappear, and you can turn the tape-recorder off.


1.10.4 *CAT The *CAT command will give a complete catalogue of all the files on a cassette. The '*' asterisk is used to distinguish the cassette operating-system commands from the BASIC commands. Rewind the tape and type: *CAT The ATOM will reply with: PLAY TAPE and you should then play the tape, and press any key to start the catalogue. As a file is encountered on the tape the filename will be printed out, together with additional information about the file: EXAMPLE XXXX XXXX XXXX XX where the 'X's represent four numbers which you can ignore for the moment (see Section 19.3 for details). When you have finished you can get back to the '>' prompt by typing CTRL (control).


1.10.5 LOAD Switch off the ATOM, in order to cause the saved text to be lost, and then switch on again and type: LOAD "EXAMPLE" The ATOM will reply with: PLAY TAPE and the tape should be rewound and played, and RETURN pressed. The computer will search through the tape for a file of the specified filename, EXAMPLE in this case, and then load it into its memory. If all is well the prompt should reappear, and then typing: LIST _9_
will give a listing of the text that was previously saved.


1.10.6 File Blocks If you save a long file on tape, and play it back, you will discover that it is broken up into a number of short blocks, with gaps in between, and that when the file is catalogued its name appears several times, once for each block. This is done for greater reliability, and if the tape is damaged in the middle of one block it will still be possible to load back the other blocks of the file. One further message that may be given when loading tapes is: REWIND TAPE This implies that you have started playing the tape in the middle of the file you wanted to load. Rewind the tape, press RETURN, and the message: PLAY TAPE will be given again.


1.10.7 Errors when Using Tape If an error is found when loading back a tape file, the message: SUM ERROR 6 is given. This might be caused by bad adjustment of the tape-recorder playback volume, a damaged or dirty tape, or recording a file over part of a previous file. If you choose an invalid name for a file, the message: NAME ERROR 118 will be given. _10_

2 Calculating in BASIC

The ATOM computer understands a language called BASIC which, because of the ease of writing programs in it, has become the most popular language for use on small computer systems. BASIC was invented in 1964 at Dartmouth College, New Hampshire, and it stands for Beginner's All-purpose Symbolic Instruction Code. This chapter introduces some of the facilities available in the BASIC language. The BASIC language consists of 'statements', 'operators', and 'functions'. The 'statements' are words like PRINT and INPUT which tell the computer what you want to do; they are followed by the things you want the computer to operate on. The 'operators’ are special symbols such as the mathematical signs '+’ and '-' meaning 'add' and ’subtract'. The 'functions' are words like the statements, but they have a numerical value; for example, RND is a function which has a random value. 2.1 PRINT This is by far the most useful BASIC statement; it enables programs to print out the results of their calculations. Try typing: PRINT 7+3 The ATOM will print: 10> The '>' prompt reappears immediately after the answer, 10, is printed out. This is the best way to use BASIC as a simple calculator; type PRINT followed by the expression you want to evaluate. Try the effect of the following: PRINT 7-3 PRINT 7*3 PRINT 7/3 You will discover that '*’ means multiply; it is the standard multiply symbol on all computers. Also '/’ means divide, but you may be surprised that the answer to 7/3 is given as 2, not 2 and 1/2. ATOM BASIC only deals in whole numbers, or integers, so the remainder after the division is lost. The remainder can be obtained by typing: PRINT 7%3 The '%' operator means ’give remainder of division'. More complex expressions are evaluated according to the standard rules of mathematics, so the expression: PRINT 2+3*4-5 has the result 9. Multiplications and divisions are performed first, followed by additions and subtractions. Round brackets can be used to I make sure that operations are performed in the correct order; anything enclosed in brackets is evaluated first. Thus the above expression could also be written: _11_
PRINT (2+(3*4))-5 There is no limit to the complexity of expressions that ATOM BASIC can evaluate, provided they will fit on two lines of the VDU. You will notice that ATOM BASIC calculates extremely rapidly. Try typing: PRINT 9*9*9*9*9*9*9*9*9 ATOM BASIC can calculate with numbers between about 2000 million and -2000 million, which gives an accuracy of between nine and ten digits. Furthermore, because whole numbers are used, all calculations in this range are exact.


2.1.1 Printing Several Things You can print the results of several calculations in one PRINT statement by separating them with commas: PRINT 7, 7*7, 7*7*7, 7*7*7*7 which will print out: 7 49 343 2401 Note that each number is printed out on the right-hand side of a column eight characters wide. This ensures that when large numbers of results are printed out they will be in neat columns on the screen. 2.1.2 Printing Strings PRINT can also be used to print out words, or indeed, any required group of characters. Arbitrary groups of characters are referred to simply as 'strings', and to identify the characters as a string they are enclosed in double quotes. For example: PRINT "THE RESULT" will cause: THE RESULT> to be printed out. The characters in quotes are copied faithfully, exactly as they appear in the PRINT statement. Thus you could type: PRINT "55*66=", 55*66 where the expression inside quotes is a string just like any other. This would print out: 55*66= 3630> 2.2 Variables - A to Z You will probably be familiar with the use of letters, such as X and N, to denote unknown quantities. E.g.: "the nth. degree", "X marks the spot", etc. In ATOM BASIC any letter of the alphabet, A to Z, may be used to denote an unknown quantity, and these are called 'variables'. The equals sign '=' is used to assign a particular value to a variable. For example, typing: X=6 will assign the value 6 to X. Now try: PRINT X and, as expected, the value of X will be printed out. Note the difference between this and: PRINT "X" _12_
The assignment statement 'X=6' should be read as 'X becomes 6' because it denotes an operation which changes the value of X, rather than a statement of fact about X. The following statement: X=X+1 is perfectly reasonable, and adds 1 to the previous value of X. In words, the new value of X is to become the old value of X plus one. Now that we can use variables to stand for numbers, they can also be used in expressions. For example, to print the first four powers of 12 we can type: T=12 PRINT T, T*T, T*T*T, T*T*T*T


2.3 Getting the Right Answer Suppose you wanted to calculate half of 777. You might type: PRINT 777/2 and you would get the answer 388. Then, to get the remainder, you would type: PRINT 777%2 and the answer will be l. So the exact answer is 388 and one half. Suppose, however, you decided to try: PRINT 1/2*777 thinking it would give 'a half times 777', you would be surprised to get the answer 0. The reason lies in tie fact that the calculation is worked out from left to right in several stages, and at every stage only the whole-number part of the result is kept. First 1/2 is calculated, and the result is 0 because the remainder is not saved. Then this is multiplied by 777 to give 0! Fortunately there is a simple rule to avoid problems like this: Do Divisions Last! The division operation is the only one that can cause a loss of accuracy; all the other operations are exact. By doing divisions last the loss of accuracy is minimised. Applying this rule to the previous example, the division by two should be done last: PRINT (1*777)/2 which is obviously the same as what was written earlier.


2.3.1 Fixed-Point Calculations An alternative way to find half of 777 is to imagine the decimal point moved one place to the right, and write: PRINT 7770/2 The result will then be 3885, or, with the decimal point moved back to the correct place, 388.5. For example, in an accounting program you would use numbers to represent pence, rather than pounds. You could then work with sums of up to 20 million pounds. Results could be printed out as follows: PRINT R/100, "POllNDS", R%100, "PENCE"


2.4 Print Field Size - '@' Numbers are normally printed out right-justified in a field of 8 character spaces. If the number needs more than 8 spaces the field _13_
size will be exceeded, and the number will be printed in full without any extra spaces. Note that the minus sign is included in the field size for negative numbers. It is sometimes convenient to alter the size of the print field. The variable '@' determines this size, and can be altered for other field widths. For example: @=32 will print one number per line, because there are 32 character positions on each line. The value of '0’ can be zero, in which case no extra spaces will be inserted before the numbers.


2.5 Printing a New Line A single quote in a PRINT statement will cause a return to the start of the next line. Thus: PRINT "A" ' "T" ' "0" ' "M" will print out: A T 0 M > This is an improvement over most other versions of BASIC, which would require four separate PRINT statements for this example.


2.6 Multiple-Statement Lines - ';' ATOM BASIC allows any number of statemerits to be strung together on each line provided they are separated by semicolons. For example the following line: A=1;B=2;C=3;PRINT A,B,C' will print: 1 2 3


2.7 Hexadecimal Numbers Numbers can also be represented in a notation called 'hexadecimal' which is especially useful for representing addresses in the computer. Hexadecimal notation is explained in section 13.1.1; all that needs to be mentioned here is that hexadecimal notation is just an alternative way of writing numbers which makes use of the digits 0 to 9 and the letters A to F. The '0' symbol, called 'hash', is used to introduce a hexadecimal number. Thus Oe9 is a perfectly good hexadecimal number (nothing to do with the variable E). PRINT #8000 will print: 32768> The PRINT statement prints the number out in decimal. 08000 is the address of the display area, and is a more convenient way of specifying this address than its decimal equivalent. A number can be printed in hexadecimal by prefixing it with an ’&' ampersand in the PRINT statement. Thus: PRINT &32768 _14_
will print: 8000>


2.8 Logical Operations In addition to the arithmetic operations already described, ATOM BASIC provides three operations called 'logical operations': '&' (AND), '\' (OR), and ':' (Exclusive-OR). These are all operations between two numbers, so there is no danger of confusing this use of '&' with its use to specify printing in hex as covered in the previous section. These are especially useful when controlling external devices from a BASIC program. Note that the '\' symbol is obtained on the keyboard by typing ’shift \', and it will appear on the display as an inverted '\' · The following table gives the results of these three operations for the numbers 0 and 1:
Operands
A B
A & B A \ B A:B
0 0 0 0 0
0 1 0 1 1
1 0 0 1 1
1 1 1 1 0
Try typing the following: PRINT 0 & 1 PRINT 1 \ 1 PRINT 1 : 1 and verify that the results agree with the table.


2.9 Peeking and Poking Many BASICs have PEEK and POKE functions which do the following: PEEK looks at the contents of a place in memory, or memory location POKE changes the contents of a memory location. The '?' operator, called 'query', is used for poking and peeking in ATOM BASIC and it provides a more elegant mechanism than the two functions provided in other BASICs. The contents of some memory location whose address is A is given by typing: PRINT ?A For example, to look at the contents of location OCOOO type: PRINT ?#COOO and the result will be 60 (this is the first location in the ATOM ROM). To change the contents of a location whose address is A to 13 just type: ?A=13 For example, to change the contents of the memory location corresponding to the top left-hand corner of the screen type: _15_
?#8000=127 and a white block will appear in the top left of the screen (see section 18.5 for an explanation). As another useful example try: ?#EI=0 which will turn the cursor off. To turn the cursor back on again type: ?#E1=#80 _16_

3 Planning a Program

The first step in writing a program, whether it will eventually be programmed in BASIC or Assembler, is to express your problem in terms of simple steps that the computer can understand. The Atom could be put to an immense number of different uses; anything from solving mathematical problems, controlling external equipment, playing games, accounting and book-keeping, waveform processing, document preparation...etc. The list is endless. Obviously all these applications cannot be included in a computer's repertoire of operations. Instead what is provided is a versatile set of more fundamental operations and functions which, in combination, can be used to solve such problems. It is therefore,up to you to become familiar with the fundamental operations that are provided, and learn how to solve problems by combining these operations into programs. Programming is rather like trying to explain to a novice cook, who understands little more than the meanings of the operations 'stir', 'boil’, etc, how to bake a cake. The recipe corresponds to the program; it consists of a list of simple operations 'stir', 'bake', with certain objects such as 'flour', 'eggs': Recipe 1. Sponge Cake 1. Mix together 4 oz. sugar and 4 oz. butter. 2. Stir in 2 eggs. 3. Stir in 4 oz. flour. 4. Put into tins. 5. Bake for 20 mins. at Mark 4. 6. Remove from oven and eat. 7. END The recipe is written so that, provided all the ingredients are already to hand, the cook can follow each command in turn without having to look ahead and worry about what is to come. Similarly, a computer only executes one operation at a time, and cannot look ahead at what is to come.


3.1 Flowcharts Before writing a program in BASIC or Assembler it is a good idea to draw a 'flowchart' indicating the operations required, and the order in which they should be performed. The generally accepted standard is for operations to be drawn inside rectangular boxes, with lines linking these boxes to show the flow of control. A simple flowchart for the program to bake a cake might be drawn as follows: _17_
atap18.gif


3.2 Decisions Many recipes do not just contain a sequence of steps to be performed, but contain conditions under which several different courses of action should be taken. For example, for a perfect cake line 5 would be better written: 5. Bake until golden brown. It would then be necessary to open the oven door every five minutes and make a decision about the colour of the cake. Decisions are represented in flowcharts by diamond-shaped boxes, with multiple exits labelled with the possible outcomes of the decision. The new flowchart would then be: _18_
atap19.gif
The action 'bake for 5 mins.' is repeatedly performed until the test 
'is it brown?' gives the answer 'yes'. Of course the program would go 
dramatically wrong if the oven were not switched on; the program would 
remain trapped in a loop.
With these two simple concepts, the action and the decision, 
almost anything can be flowcharted. Part of the trick in flowcharting 
programs is to decide how much detail to put into the flowchart. For 
example, in the cake program it would be possible to add the test 'is 
butter and sugar mixed?' and if not, loop back to the operation 'mix 
butter and sugar'. Usually flowcharts should be kept as short as 
possible so that the logic of the program is not obscured by a lot of
unnecessary fine detail.


3.3 Counting Recipes sometimes require a particular series of operations to be performed a fixed number of times. The following recipe for puff pastry illustrates this: Recipe 2. Puff Pastry 1. Mix 6 oz. flour with 2 oz. butter. 2. Roll out pastry. 3. Spread with 2 oz. butter. _19_
4. Fold in half. 5. Repeat steps 2 to 4 a further 3 times. 6. END In this recipe the cook has to perform operations a total of 4 times. A cook would probably keep a mental note of how many times he has performed these operations; for the sake of the flowchart it is convenient to give the number of operations a label, such as T. The flowchart of the puff pastry recipe would then be:
atap20.gif
The loop consisting of statements 2 to 4 is performed 4 times; the 
test at the end gives the answer 'no' for T=l, 2, and 3, and the 
answer 'yes' for T=4.
To perform an operation several times in a BASIC or Assembler 
program an identical method can be used; a counter, such as T, is used 
to count the number of operations and the counter is tested each time 
to determine whether enough operations have been completed.


3.4 Subroutines A recipe may include a reference to another recipe. For example, a typical recipe for apple tart might be as follows: _20_
Recipe 3. Apple Tart 1. Peel and core 6 cooking apples. 2. Make pastry as in recipe 2. 3. Line tart tin with pastry. 4. Put in apple. 5. Bake for 40 mins. mark 4. 6. END To perform step 2 it is necessary to insert a marker in the book at the place of the original recipe, find the new recipe and follow it, and then return to the original recipe and carry on at the next statement. In computer programming a reference to a separate routine is termed a 'subroutine call'. The main recipe, for apple tart, is the main routine; one of its statements calls the recipe for puff pastry, the subroutine. Note that the subroutine could be referred to many times throughout the recipe book; in the recipe for steak and kidney pie, for example. One reason for giving it separately is to save space; otherwise it would have to be reproduced for every recipe that needed it. Note that, in order not to lose his place, the cook needed a marker to insert in the recipe book so that he should know where to return to at the end of the subroutine. In BASIC or assembler programs the computer keeps a record of where you were when you call a subroutine, and returns you there automatically at the end of the subroutine. In other respects, the process of executing a subroutine on a computer is just like this analogy.


3.5 Planninq a Program Before writing a program in BASIC or Assembler it is a good idea to express the problem in one of the forms used in this chapter: l. As a list of numbered steps describing, in words, exactly what to do at each step. 2.As a flowchart using the following symbols: atap21.gif Having done this, the job of writing the program in BASIC or Assembler will be relatively easy. _21_
_22_

4 Writing a BASIC Program

Commands and statements typed after the ATOM's prompt are executed immediately, as we have seen in Chapter 3. However if you start the line with a number, the line is not executed but stored as text in the ATOM's memory.


4.1 RUN First type 'NEW' to clear the text area. Then try typing in the following: 10 PRINT "A PROGRAM!" 20 END When these lines have been typed in you can list the text by typing LIST. Now type: RUN The stored text will be executed, one statement at a time, starting with the lowest-numbered statement, and the message 'A PROGRAM!' will be printed out. The text you entered formed a 'program', and the program was executed, statement by statement, when you typed RUN. The END statement is used to stop execution of the program; if it is ommitted an error message will be given.


4.2 INPUT Type NEW again, and then enter the following program: 10 INPUT N 20 N=N+1 30 PRINT N The INPUT statement enables you to supply numbers to a running program. When it is executed it will print a question mark and wait for a number to be typed in. The variable specified in the INPUT statement will then be set to the value typed in. To illustrate, type: RUN The program will add 1 to the number you type in; try running it again and try different numbers. The INPUT statement may contain more than one variable; a question mark will be printed for each one, and the values typed in will be assigned to the variables in turn. The INPUT statement may also contain strings; these will be printed out before each question mark. The following program illustrates this; it converts Farenheit to Celsius (Centigrade), giving the answer to the nearest degree: 10 INPUT "FARENHEIT" F 20 PRINT (10*F-315)/18 "CELSIUS" 30 END _23_
The value, in Farenheit, is stored in the variable F. The expression in the PRINT statement converts this to Celsius.


4.3 Comments - REM The REM statement means 'remark’; everything on the line following the REM statement will be ignored when the program is being executed, so it can be used to insert comments into a program. For example: 5 REM PROGRAM FOR TEMPERATURE CONVERSION


4.4 Functions Functions are operations that return a value. Functions are like statements in that they have names, consisting of a sequence of letters, but unlike statements they return a value and so can appear within expressions.


4.4.1 RND The RND function returns a random number with a value anywhere between the most negative and most positive numbers that can be represented in BASIC. To obtain smaller random numbers the '%' remainder operator can be used; for example: PRINT RND%4 will print a number between -3 and +3.


4.4.2 TOP TOP returns the address of the first free memory location after the BASIC program. PRINT &TOP will print TOP in hexadecimal. This will be 48202 if you have not entered a program (or have just typed NEW) on the unexpanded ATOM, and #2902 on an expanded ATOM. PRINT TOP-#8200 is a useful way of finding out how many bytes are used up by a program; on an unexpanded ATOM there is a total of 512 bytes for programs.


4.4.3 ABS The ABS function can be used to give the absolute or positive value of a number; the number is written in brackets after the function name. For example: PRINT ABS(-57) will print 57. One use of ABS is in generating positive random numbers. For example: PRINT ABS(RND)%6 gives a random number between 0 and 5.


4.5 Escape - ESC It is possible to create programs which will never stop; see the following example in section 4.6. The escape key 'ESC' at the top left of the keyboard will stop any BASIC program and return control to the '>' prompt. _24_
4.6 GOTO In the above programs the statements were simply executed in ascending order of their line numbers. However it is sometimes necessary to transfer control forwards or backwards to somewhere other than the next numbered statement. The GOTO (go to) statement is used for this purpose; the GOTO statement specifies the statement to be executed next. For example, type: 1 REM Stars 10 PRINT "*" 20 GOTO 10 A flowchart for this program makes it clear that the program will never stop printing stars: atap25.gif To stop the program you will have to type ESC (escape). 4.6.1 Labels - a to z ATOM BASIC offers another option for the GOTO statement. Instead of giving the number of the statement to be executed next, a statement can be designated by a 'label', and the GOTO is followed by the required label. A label can be one of the lower-case letters a to z, which are obtained on the ATOM by typing the letter with the shift key held down. Labels appear on the VDU as upper-case inverted letters, so they are very easily identified in programs. For typographical convenience labels will be represented as lower-case letters in this manual. To illustrate the use of labels, rewrite the ’STARS' program as follows, using the label 's': 10s PRINT "*" 20 GOTO s Note that there must be no spaces between the line number and the label. There are two advantages to using labels, rather than line numbers, in GOTO statements. First, programs are clearer, and do not depend on how the program lines are numbered. Secondly, the GOTO statement is faster using a label than using a line number. To demonstrate this, enter the following program which generates a tone of 187 Hz in the loudspeaker: 10 P=#B002 20a ?P=?P:4; GOTO a This program works as follows: P is the location corresponding to the input/output port, and exclusive-ORing this location with 4 will change the output line connected to the loudspeaker. The frequency generated implies that the statements on line 20 are executed in about 2.5 milliseconds (twice per cycle). _25_
Try removing the label and rewrite the program as follows: 10 P=OB002 20 ?P=?P:4; GOTO 20 The GOTO statement is now slightly slower, and the tone generated will have the lower frequency of 144 Hz. The highest frequency that can be generated by a BASIC program is 322 Hz, as follows: 10 REM 322 Hz 20 P=#B002 30 FOR Z=O TO 10000000 STEP 4;?P=Z;N. To play tunes you will need to use an assembler program; see Section 15.4.


4.6.2 Switches The GOTO statement may be followed by any expression which evaluates to a valid line number; for example: 10 REM Two-Way Switch 20 INPUT "TYPE 1 OR 2" L 30 GOTO (40*L) 40 PRINT "ONE" 50 END 80 PRINT "TWO" 90 END If L is 1 the expression (40*L) will be equal to 40, and the program will print 'ONE'. If L is 2 the expression will be equal to 80 and the program will print 'TWO’. The flowchart for this program is as follows: atap26.gif


4.6.3 Multi-Way Switches Finally here is an example of a multi-way switch using GOTO. The program calculates a random number between 0 and 5 and then goes to a _26_
line number between 30 and 35. Each of these lines consists of a PRINT statement which prints' the face of a dice. The single quote in the print statement gives a 'return' to the start of the next line. 10 REM Dice Tossing 20 GOTO (30+ABS(RND)%6) 30 PRINT'" *"''; END 31 PRINT" *"''"*"'; END 32 PRINT" *"'" *”’”*"'; END 33 PRINT"* *”''"* *"’; END 34 PRINT"* *"'" *"'”* *"'; END 35 PRINT"* *"'"* *"'"* *"'; END Description of Program: 20 Choose random number between 30 and 35 30-35 Print corresponding face of a dice Sample runs: >RUN * * * >RUN * * * * * >RUN * * * * * *


4.7 Graphics The ATOM has no less than 9 different graphics modes available from BASIC. This section provides a brief introduction to graphics mode 0, the lowest resolution mode, which is available on the unexpanded ATOM. With more memory added to the ATOM the other graphics modes are available, and these are explained in Chapter 11. A special feature of mode 0 is that it is possible to mix graphics with any of the ATOM's characters. Graphics treats the screen as a piece of graph paper on which it is possible to draw lines and plot points. Points on the screen are called 'picture elements' or ’pixels' for short, because they are actually small squares. Each pixel on the screen is specified by its coordinates in the two directions, horizontal and vertical, and these coordinates will be referred to as X and Y respectively. The graphics screen is labelled as follows in mode 0:
0,47   63,47
^
|
Y
|
atom_1.png  
0,0 -- X --> 63,0
_27_


4.7.1 CLEAR To prepare the screen for graphics the statement CLEAR is used. It is followed by the graphics mode number. On the unexpanded ATGN the only legal option is: CLEAR 0


4.7.2 MOVE Any point on the screen can be specified by moving the 'graphics cursor' to that point with the MOVE statement. The graphics cursor does not show on the screen, and it is different from the ordinary cursor which is visible in character mode. The format of the statement is: MOVE X,Y where X and Y can be numbers, or arbitrary expressions provided they are enclosed in brackets. For example, to move the graphics cursor to the origin, X=0 Y=O, type: MOVE 0,0 The MOVE statement will normally be the first graphics statement of any program.


4.7.3 DRAW The DRAW statement will plot a line anywhere on the screen. The line starts from the position of the graphics cursor, and ends at the point specified in the statement, and tie graphics cursor will be moved to that point. For example: DRAW 63,47 will draw a line to the top right-hand corner of the screen, and leave the graphics cursor at that point. It is quite legal, and safe, to draw off the screen; the line will just not appear.


4.7.4 Example The following simple program will draw a rectangle, rotated by the amount entered for R. Try typing in numbers between 0 and 47 for R: 10 REM Rotating Rectangle 20 x=63; Y=47 30 INPUT R 40 CLEAR 0 50 MOVE R,O 60 DRAW X,R; DRAW (X-R),Y 70 DRAW 0,(Y-R); DRAW R,O 80 GOTO 30


4.7.5 Plotting Points One way of plotting a single point at X,Y on the screen is to write: MOVE X,Y; DRAW X,Y A more elegant way is given in Section 11.3.


4.8 Conditions - IP...THEN One of the most useful facilities in BASIC is the ability to execute a statement only under certain specified conditions. To do this the IF...THEN statement is used; for example: _28_
IF A=0 THEN PRINT "ZERO" will execute the PRINT statement, and print "ZERO", only if the condition A=0 is true; otherwise everything after THEN will be skipped and execution will continue with the next line.


4.8.1 Relational Operators The part of the IF...THEN statement after the IF is the 'condition' which can be any two expressions separated by a 'relational operator' which compares the two expressions. Six different relational operators can be used: = equal << not equal < greater than <="less than or equal < less than"<= greater than or equal where each operator on the left is the opposite of the operator on the right. The expressions on each side of the relational operators can be as complicated as required, and the order is unimportant. There is no need to put brackets around the expressions. For example, the following program prints one of three messages depending on whether a number typed in is less than 7, equal to 7, or greater than 7: 10 REM Guess a number 20 INPUT"GUESS A NUMBER" N 30 IF N<7 then print ntoo small" 40 if n="7 THEN PRINT CORRECT! 50 IF N">7 THEN PRINT "TOO LARGE" 60 END A flowchart for this program is as follows: _29_
atap30.gif 4.8.2 THEN Statement The statement after THEN can be any statement, even an assignment statement as in: IF A=7 THEN A=6 Note that the meaning of each '=' sign is different. The first 'A=7' is a condition which can be either true or false; the second 'A=6' is an assignment statement which instructs the computer to set the variable A to the value 6. To make this distinction clear the above statement should be read as: 'If A is equal to 7 then A becomes 6'. 4.8.3 Conjunctions - AND and OR Conditions can be strung together using the conjunctions AND and OR, so, for example: 10 INPUT A,B 20 IF A=2 AND B=2 THEN PRINT "BOTHI 30 GOTO 10 _30_
will only print "BOTH" if both A and B are given the value 2. Alternatively: 10 INPUT A,B 20 IF A=2 OR B=2 THEN PRINT ”EITHER" 30 GOTO 40 will only print "EITHER" if at least one of A and B is equal to 2. 4.9 Logical Variables An alternative form for the condition in an IF...THEN statement is to specify a variable whose value denotes either 'true' or 'false'. The values 'true’ and 'false' are represented by 1 and 0 respectively, so: A=1; B=0 sets A to 'true' and B 'false'. Logical variables can be used in place of conditions in the IF statement; for example: IF A THEN PRINT "TRUE" will print "TRUE". A logical variable can also be set to the value of a condition: A=(L=100) This statement will set A to 'true' if L is 100, and to 'false’ otherwise. The condition must be placed in brackets as shown. 4.10 Iteration One way of printing the powers of 2 would be to write: 10 REM Powers of Two 20 P=1; T=2; @=0 30 PRINT "2 ", P, "= ", T ' 40 T=T*2; P=P+1 50 GOTO 30 which will print out: 2 ^ 1 = 2 2 ^ 2 = 4 2 ^ 3 = 8 2 ^ 4 = 16 2 ^ 5 = 32 2 ^ 6 = 64 and so on without stopping. This is a bit inelegant; suppose we wished to print out just the first 12 powers of 2. It is simply a matter of detecting when the 12th. power has just been printed out, and stopping then. This can be done with the IF statement as follows: 10 REM First Twelve Powers of Two 20 P=l; T=2; @=0 30 PRINT "2" ^, P, " = ", T ' 40 T=T*2; P=P+1 50 IF PC=12 GOTO 30 60 END The IF statement is followed by a GOTO statement; if P is less than 12 the condition will be true, and the program will go back to line 30. _31_
After the twelth power of 2 has been printed out P will have the value 13, which is not less than or equal to 12, and so the program will stop. With the IF statement we have the ability to make the computer do vast amounts of work as a result of very little effort on our part. For example we can print,256 exclamation marks simply by running the following program: 10 I=0 20 PRINT"!"; I=I+1 30 IF I<256 goto 20 40 end


4.10.1 Cubic Curve Perhaps a more useful example is the following program, which calculates the value of: x^3 - 600x for 64 values of x and plots a graph of the resulting curve: 1 REM Cubic Curve 10 CLEAR 0 20 MOVE 0,24; DRAW fi3,24 30 MOVE 32,0; DRAW 32,47 40 MOVE -1,-1 50 X=-33 55 Y=(X*X*X-600*X)/400 60 DRAW (32+X),(24+Y) 70 X=X+1 80 IF X<33 then goto 55 90 end description of program: 10 use graphics mode 0 20-30 draw axes 40 move graphics cursor off screen 50-80 plot curve for values of x from -32 to 32 55 equation to be evaluated divided by 400 to bring the interesting part of the cubic curve into range 60 draw to next point, with origin at (32,24). program size: 190 bytes _32_

5 Loops

The previous section showed how the IF statement could be used to cause the same statements to be executed several times. Recall the program: 10 I=0 20 PRINT"!"; I=I+1 30 IF I<256 goto 20 40 end which prints out 256 exclamation marks (half a screen full). this iterative loop is such a frequently-used operation in basic that all basics provide a special pair of statements for this purpose, and atom basic provides a second type of loop for greater flexibility.


5.1 FOR...NEXT Loops The FOR statement, together with the NEXT statement, causes a set of statements to be executed for a range of values of a specified variable. To illustrate, the above example can be rewritten using a FOR...NEXT loop as follows: 10 FOR I=1 TO 256 20 PRINT "!" 30 NEXT I 40 END The FOR statement specifies that the statements up to the matching NEXT statement should be executed for each value of I from 1 to 256 (inclusive). In this example there is one statement between the FOR and NEXT statements, namely: PRINT "!" This statement has been indented in the program to make the loop structure clearer; in fact the spaces are ignored by BASIC. The NEXT statement specifies the variable that was specified in the corresponding FOR statement. This variable, I in the above example, is called the 'control variable' of the loop; it can be any of the variables A to Z. The value of the control variable can be used inside the loop, if required. To illustrate, the following program prints out all multiples of 12 up to 12*12: 10 FOR M=1 TO 12 20 PRINT M*12 30 NEXT M 40 END The range of values specified in the FOR statement can be anything you wish, even arbitrary expressions. Remember, though, that the loop is always executed at least once, so the program: _33_


10 FOR N=1 TO 0 20 PRINT N 30 NEXT N 40 END will print '1' before stopping.


5.1.1 STEP Size It is also possible to specify a STEP size in the FOR statement; the STEP size will be added to the control variable each time round the loop, until the control variable exceeds the value specified after TO. If the STEP size is omitted it is assumed to be 1. This provides us with an alternative way of printing the multiples of 12: 10 FOR M=12 TO 12*12 STEP 12 20 PRINT M 30 NEXT M 40 END


5.1.2 Graph Plotting Using FOR...NEXT The FOR...NEXT loop is extremely useful when plotting graphs using the ATOM's graphics facilities. Try rewriting the Cubic Curve program of Section 4.10.1 using a FOR...NEXT loop. The following curve-stitching program is quite fun, especially in the higher graphics modes. It simulates the curves produced by stitching with threads stretched between two lines of holes in a square of cardboard. The curve produced as the envelope of all the threads is a parabola: 1 REM Curve Stitching in a Square 10 V=46 20 INPUT Q 30 CLEAR 0 40 FOR Z=O TO V STEP Q; Y=V-Z 50 MOVE O,Z; DRAW Y,O 60 MOVE Y,V; DRAW V,Z 70 NEXT Z 80 END The value of Q typed in should be between 2 and 9 for best results; V determines the size of the square that is drawn. The program works best when V is a multiple of Q.


5.2 DO...UNTIL Loops ATOM BASIC provides an alternative pair of loop-control statements: DO and UNTIL. The UNTIL statement is followed by a condition, and everything between the DO statement and the UNTIL statement is repeatedly executed until the condition becomes true. So, to print 256 exclamation marks in yet another way write: 10 I=0 20 DO 30 I=I+1 40 PRINT "!" 50 UNTIL I=256 50 END Again, the statements inside the DO...UNTIL loop may be indented to make the structure clearer. _34_
The DO...UNTIL loop is most useful in cases where a program is to carry on until certain conditions are satisfied before it will stop. To illustrate, the following program prompts for a series of numbers, and adds them together. When a zero is entered the program terminates and prints out the sum: 10 S=O 20 DO INPUT J 30 S=S+J 40 UNTIL J=O 50 PRINT "SUM =", S 60 END Note that a statement may follow the DO statement, as in this example.


5.2.1 Greatest Common Divisor The following simple program uses a DO...UNTIL loop in the calculation of the greatest common divisor (GCD) of two numbers; i.e. the largest number that will divide exactly into both of them. For example, the GCD of 26 and 65 is 13. If the numbers are coprime the GCD will be l. 1 REM Greatest Common Divisor 80 INPUT A,B 90 DO A=ASB 100 IFABS(B)>ABS(A) THEN T=B; B=A; A=T 120 UNTIL B=O 130 PRINT "GCD =" A ' 140 END Description of Program: 80 Input two numbers 90 Set A to remainder when it is divided by B 100 Make A the larger of the two numbers 120 Stop when B is zero 130 A is the greatest common divisor. Variables: A,B - Numbers T - Temporary variable Program size: 137 bytes The method is known as Euclid’s algorithm, and to see it working insert a line: 95 PRINT A,B' The ABS functions ensure that the program will work for negative, as well as positive, numbers.


5.2.2 Successive Approximation The DO...UNTIL loop construction is especially useful for problems involving successive approximation, where the value of a function is calculated by obtaining better and better approximations until some criterion of accuracy is met. The following iterative program calculates the square root of any number up to about 2,000,000,000. Also shown is the output obtained when calculating the square root of 200,000,000: 10 REM Square Root 20 INPUT S _35_
100 Q=s/2 110 DO Q=(Q+S/Q)/2 120 UNTIL (Q-1)*(Q-1)<S AND (Q+1)*(Q+1)>S 130 PRINT Q 140 END Description of Program: 20 Input number 100 Choose starting value 110 Calculate next approximation 120 Carry on until the square lies between the squares of the numbers either side of the root. 130 Print square root. Variables: Q - Square root S - Number Program size: 118 bytes Sample run: >RUN ?200000000 14142>


5.3 Nested Loops FOR...NEXT and DO...UNTIL loops may be nested; the following example will print the squares, cubes, and fourth powers of the numbers 1 to 15 in a neat table: 1 REM Powers of Numbers 5 PRINT" X X^2" 8 PRINT" X^3 X^4" 10 FOR N=1 TO 15 20 J=N 30 FOR M=1 TO 4 40 PRINT J; J=J*N 50 NEXT M 60 NEXT N 70 END The statements numbered 20 to 50 are executed 15 times, for every value of N from 1 to 15. For each value of N the statements on line 40 are executed four times, for values of M from 1 to 4. Thus 15*4 or 60 numbers are printed out.


5.3.1 Mis-Nested Loops Note that loops must be nested correctly. The following attempt at printing out 100 pairs of numbers will not work: 10 FOR A=1 TO 10 20 FOR B=1 TO 10 30 PRINT A,B 40 NEXT A 50 NEXT B The program will, if RUN, give an error (ERROR 230). The reason for the error will become clear if you try to indent the statements within each loop, as in the previous example. _36_


5.4 WAIT Statement ATOM BASIC includes an accurate timing facility, derived from the main CPU clock. To understand the operation of the WAIT statement, imagine that the ATOM contains a clock which 'ticks' sixty times a second. The WAIT statement causes execution to stop until the next clock tick. Thus it automatically synchronises the program to an accurate time. The WAIT statement makes it a simple matter to write programs to give any required delay. For example, the following program gives a delay of 10 seconds: FOR N=1 TO 10*60; WAIT; NEXT N You are perhaps wondering why WAIT does not just give a delay of 1/60 second, rather than waiting for the next clock tick. The reason is that if only a delay function were provided, you would have to know exactly how long the other statements in the loop took to execute if you wanted accurate timinq. In fact, with the WAIT function, all you need to do is to ensure that the statements in the loop take less than 1/60th. of a second, so as not to miss the next tick.


5.4.1 Digital Clock The following digital clock displays the time as six digits in the top left-hand corner of the screen. 10 REM Digital Clock 20 INPUT "TIME" H,MgS 30 PRINT $12; ?#El=0 40 T=((H*100)+M)*100+S 50 DO FOR S=l TO 55; WAIT; NEXT S 60 PRINT $30,T; T=T+1 70 IF T%100=60 THEN T=T+40 80 IF T%10000=6000 THEN T=T+4000 90 UNTIL 0 Description of Program: 20 Input the starting time 30 Clear screen; turn off cursor 40 Set up time as 6-digit number 50 Use up rest of a second 60 Print T in top left-hand corner of screen 70-80 Update minutes and hours Variables: H - Hours M - Minutes S - Seconds T - Six-digit number representing time Program size: 216 bytes To turn the cursor back on after running this program type a form-feed; i.e. CTRL-L.


5.4.2 Reaction Timer The following reaction-timer program uses WAIT to calculate your reaction time, and prints out the time in centiseconds (i.e. hundredths of a second) to the nearest 2 centiseconds. It blanks the screen, and then, after a random delay, displays a dot at a random place on the screen. When you see the dot you should press the SHIFT key as quickly as possible; the program will then display your reaction time. _37_
1 REM Reaction Timer 10 CLEAR 0 20 X=ABS(RND)%64; Y=ABS(RND)%48 30 FOR N=l TO ABS(RND)%600+300 40 WAIT; NEXT N 50 MOVE X,Y; DRAW X,Y 60 T=O 70 DO T=T+1; WAIT 80 UNTIL ?#Bool<>#FF 90 PRINT "REACTION TIME =" 100 PRINT T*10/6, "CSEC." 110 IF T>18 PRINT "WAKE UP!" 120 END Description of Program: 20 Choose random X,Y coordinates for point on screen. 30-40 Wait for random time between 6 and 9 seconds. 50 Plot point at X,Y 60-70 Count sixtieths of a second 80 #b001 is the address of the input port to which the SHIFT key is connected; the contents of this location are #ff until the SHIFT key is pressed. 90-100 Print reaction time converted to centiseconds. 110 If appalling reactions, print message. Variables: N - Counter for random delay T - counter in sixtieths of a second for reaction time X,Y - random coordinates for point on screen. Program size: 273 bytes _38_

6 Subroutines

As soon as a program becomes longer than a few lines it is probably more convenient to think of it as a sequence of steps, each step being written as a separate 'routine', an independent piece of program which can be tested in isolation, and which can be incorporated into other programs when the same function is needed.


6.1 GOSUB Sections of program can be isolated from the rest of the program using a BASIC construction called the 'subroutine'. In the main program a statement such as: GOSUB 1000 causes control to be transferred to the statement at line 1000. The statements from line 1000 comprise the subroutine. The subroutine is terminated by a statement: RETURN which causes a jump back to the main 'calling' program to the statement immediately following the GOSUB 1000. It is just as if the statements from 1000 up to the RETURN statement had simply been inserted in place of the GOSUB 1000 statement in the main program. As an example, consider the following program: 10 A=10 20 GOSUB 100 30 A=20 40 GOSUB 100 50 END 100 PRINT A ' 110 RETURN Lines 100 and 110 form a subroutine, separate from the rest of the program, and they are terminated by RETURN. The subroutine is called twice from the main program, in lines 20 and 40. The program, when RUN, will print:


6.1.1 Chequebook-Balancing Program As a more serious example, consider a program for balancing a chequebook. The program will have three distinct stages; reading in the credits, reading in the debits, and printing the final amount. We can immediately write the main program as: 10 REM Chequebook-Balancing Program 20 PRINT "ENTER CREDITS"' 30 GOSUB 1000 _39_
40 PRINT ”ENTER DEBITS"' 50 GOSUB 2000 60 PRINT "TOTAL IS" 70 GOSUB 3000 80 END Now all we have to do is write the subroutines at lines 1000, 2000, and 3000! The subroutines might be written as follows: 1000 REM Sum Credits in C 1010 REM Changes C,J 1020 C=O 1030 DO INPUT J; C=C+J 1040 UNTIL J=O 1050 RETURN 2000 REM Sum Debits in D 2010 REM Changes D,J 2020 D=O 2030 DO INPUT J; D=D+J 2040 UNTIL J=O 2050 RETURN 3000 REM Print Total in T 3010 REM Changes T; Uses C,D 3020 T=C-D; 0=5 3030 PRINT T/100," POUNDS",T%100," PENCE" 3040 RETURN Values are entered in pence, and entering zero will terminate the list of credits or debits. The two subroutines at 1000 and 2000 are strikingly similar, and this suggests that it might be possible to dispense with one of them. Indeed, the main part of the chequebook-balancing program can be written as follows, eliminating subroutine 1000: 10 REM Chequebook-Balancing Program 20 PRINT "ENTER CREDITS" 30 GOSUB 2000 40 C=D 50 PRINT "ENTER DEBITS" 60 GOSUB 2000 70 PRINT "TOTAL IS" 80 GOSUB 3000 90 END In conclusion, subroutines have two important uses: 1. To divide programs into modules that can be written and tested separately, thereby making it easier to understand the operation of the program. 2. To make it possible to use the same piece of program for a number of similar, related, functions. As a rough guide, if a program is too long to fit onto the screen of the VDU it should be broken down into subroutines. Each subroutine should state clearly, in REM statements at the start of the subroutine, the purpose of the subroutine, which variables are used by the subroutine, and which variables are altered by the subroutine. A few moments spent documenting the operation of the subroutine in this _40_
way will save hours spent at a later date trying to debug a program which uses the subroutine.


6.2 GOSUB Label The GOSUB statement is just like the GOTO statement that has already been described, in that it can be followed by a line number, an expession evaluating to a line number, or a label. Labels are of the form a to z, and the first line of the subroutine should contain the label immediately following the line number.


6.2.1 Linear Interpolation The following program uses linear interpolation to find the roots of an equation using only integer arithmetic, although the program could be modified to use floating-point statements. The equation is specified in a subroutine, y, giving Y in terms of X; the program finds solutions for Y=O. As given, the program finds the root of the equation: x2 - x - 1 = 0 The larger root of this equation is phi, the golden ratio. A scaling factor of S=1000 is included in the equation so that calculations can be performed to three decimal places. The program prompts for two values of X which lie either side of the root required. 1 REM Linear Interpolation 5 S=1000; 0=0; I=1 10 INPUT "X1",A,"X2",B 20 A=A*S; B=B*S 30 X=A; GOSUB y; C=Y 40 X=B; GOSUB y; D=Y 50 IF C*D<0 goto 80 60 print "root not bracketed" 70 end 80 do i="I+1 90" x="B-(B-A)*D/(D-C); GOSUB y 100 IF C*Y<0 THEN" a="X;" c="Y; GOTO 120 110" b="X;" d="Y 120 UNTIL ABS(A-B)<2 OR ABS(Y)<2 130 PRINTROOT IS" x="140 IF X<0 PRINT - 145 PRINT ABS(X)/S,. 150 DO" x="ABS(X)%S;" s="S/10 155 PRINT X/S; UNTIL" s="1 160 PRINT'NEEDED ,I, ITERATIONS.' 170" end 200yy="X*X/S-X-1*S 210 RETURN Description of Program: 5-70 Check that starting values bracket a root 80-120 Find root by successive approximation 130-145 Print integer part of root 150-155 Print decimal places 160 Print number of iterations needed 200-210 y: Subroutine giving Y in terms of X, with appropriate scaling. _41_

Variables: A - Lower starting value of X B - Upper starting value of X C - Value of Y for X=A D - Value of Y for X=B I - Iteration number S - Scaling factor; all numbers are multiplied by S and held as integers. X - Root being approximated Y - Value of equation for given,X Program size - 466 bytes Sample run: >RUN X1?1 X2?3 ROOT IS X= 1.618 NEEDED 7 ITERATIONS.


6.3 Subroutines Calling Subroutines Often the task carried out by a subroutine may itself usefully be broken down into a number of smaller steps, and so it might be convenient to include calls to subroutines within other subroutines. This is perfectly legal, and subroutines may be nested up to a maximum depth of 15 calls.


6.4 Recursive Subroutine Calls Sometimes a problem can be more simply expressed if it is allowed to include a reference to itself. When a subroutine includes a call to itself in this way it is known as a 'recursive' subroutine call, and it is possible to use recursive calls in ATOM BASIC provided that the depth of recursion is limited to 15 calls. The following half-hearted program uses a recursive call to print out ten stars without using a loop: 10 REM Recursive Stars 20 P=10; GOSUB p 30 END 100pREN Print P stars 110 IF P=O RETURN 120 P=P-1 130 GOSUB p; REM Print P-1 stars 140 PRINT "*" 150 RETURN This program could, of course, be written more effectively using a simple FOR...NEXT loop. The following programs, however, use recursion to great benefit to solve mathematical problems that would be much harder to solve using iteration alone.


6.4.1 Tower of Hanoi Problem In the Tower of Hanoi problem three pegs are fastened to a stand, and there are a number of wooden discs each with a hole at its centre. The discs are all of different diameters, and they all start on one peg, arranged in order of size with the largest disc at the bottom of the pile. The problem is to shift the pile to another peg by transferring _42_
one disc at a time, with the restriction that no disc may be placed on top of a smaller disc. The number of moves required rises rapidly with the number of discs used; the problem was classically described with 64 discs, and moving one disc per second the solution of this problem would take more than 500,000 million years! A recursive solution to the problem, stated in words, is: To move F discs from peg A to peg B: 1. Move F-1 discs to peg C. 2. Move bottom disc to peg B. 3. Move F-1 discs to peg B. Also, when F is zero there is no need to do anything. Steps 1 and 3 of the procedure contain a reference to the whole procedure, so the solution is recursive. The following program will solve the problem for up to 13 discs, and displays the piles of discs at every stage in the solution: 1 REM Tower of Hanoi 10 PRINT$12 20 A=TOP;D=A+4 40 V=-3;W=-1 60 !D=$1020300;!A=0 70 INPUT"NUMBER OF DISCS "F 80 A?1=F;?D=F 85 N=64/3 90 CLEARO 100 FORQ=1TOF;MOVE(F-Q),(2*(F-Q));PLOTlg(2*Q-1),0;NEXT 110 GOSUBh;END 1000hIF?D=O RETURN 1010 D!4=!D-1;D?6=D?1;D?5=D?2;D=D+4;GOSUBh 1020 MOVE(F-D?-4+D?V*N-N),(D?V?A*2);PLOT1,(D?-4*2-1),0 1030 MOVE(D?W*N-N),(D?W?A*2-2);PLOT3,(F+D?-4),0 1040 A?(D?W)=A?(D?W)+W;A?(D?V)=A?(D?V)-W 1050 D?3=D?-2;D?2=D?W;D?1=D?V;GOSUBh 1060 D=D-4;RETURN Description of Program: 100 Draw starting pile of discs 110 Subroutine h is called recursively to move the number of discs specified in ?D. 1000 h: Subroutine to move ?D discs 1010 Recursive call to move ?D-1 discs 1020 Draw new disc on screen 1030 Remove old disc from screen 1040 Set up array A 1050 Recursive call to put back ?D-1 discs Variables: A?N - Number of discs on pile N D - Stack pointer ?D - How many discs to transfer D?1 - Destination Pile D?2 - Intermediate pile D?3 - Source pile F - Total number of discs N - One third of screen width V - Constant W - Constant Program size: 461 bytes Stack usage: (4 * number of discs) bytes _43_


6.3.2 Eight Queens Problem A classical mathematical problem consists of placing eight queens on a chessboard so that no queen attacks any other. The following program find all possible solutions to the problem, and draws a diagram of the board to show each solution as it is found. The program uses many abbreviations to keep it small enough to fit on an unexpanded ATOM (for a complete explanation of these abbreviations, see section 10.1): 1 REM Eight Queens 30 C=O;D=TOP;E=D+3;A=D+27;!D=O 60 0=0;GOS.t;P.$13"THERE ARE "C" SOLUTIONS"';END 100tIF?D=OFF C=C+1;GOTOd 110 ?A=(?D\D?1\3?2):#FF 120lIF?A=OR. 130 A?1=?A&-?A 140 ?E=?D\A?1;E?l=(D?1\A?1)*2;E?2=(D?2\A?1)/2 150 D=D+3;E=E+3;A=A+2;GOS.t;D=D-3;E=E-3;A=A-2 160 ?A=?A&(A?1:OFF);GOT01 200dCLEARO;FORZ=OT032S.4;MOVEO,Z;DRAW31,Z;MOVEZ,O;DRAWZ,32;N. 210 Q=o;forz=3t024step3;p=top?z-Q;s=-2;dos=s+4;p=p/2;untilp=0 220 Q=top?z;plot13,(2/3*4-2),s;n.;p.$30 C;R. Description of Program: 30 Initialise array space. D is vector of attacks, ?D is row attacks, D?1 is left diagonal attacks, D?2 is right diagonal attacks. 60 Call recursive analyser and print answer. 100 t: Recursive analyser: if all rows attacked have found a solution. 110 Calculate possible places to put new queen. 120 If no possible place, end this recursive attempt. 130 Find least significant bit in possible places to use as new queen position. 140 Calculate new attacked values. 150 Recursive call of analyser. 160 Remove this position from possible position and see if done. 200 d: Have solution, display board matrix. 210 Plot pixels at positions of queens. 220 Print the solution number at screen top and end recursion. Variables: ?A - Possible position; value of A changes C - Solutions counter ?D - Row attacks; value of D changes E - Holds D+3 to make program shorter Program size: 440 bytes Vectors: 30 bytes Total storage: 470 bytes. _44_

7 Arrays and Vectors

So far we have met just 26 variables, called A to Z. Suppose you wanted to plot a graph showing the mean temperature for every month of the year. You could, at a pinch, use the twelve letters A to L to represent the mean temperatures, and read in the temperatures by saying: INPUT A,B,C,D,E,F,G,H,I,J,K,L However there is a much better way. A mathematician might call the list of temperatures by the names: t1, t2, t3, ..... t12. where the 'subscript', the number written below the- line, is the number of the month in the year. This representation of the twelve temperatures is much more meaningful than using twelve different letters to stand for them, and there is no doubt about which symbol represents the temperature of, for example, the third month. A similar series of variables can be created in ATOM BASIC, and these are called 'arrays'. Each array consists of an array 'identifier', or name, corresponding to the name 't' in the above example, and a 'subscript'. On most computers there is no facility for writing subscripts, so some other representation is used. Each member of the array can act as a completely independent variable, capable of holding a value just like the variables A to Z. The members of an array are called the array 'elements'. The total number of possible elements depends on how the array was set up; in the above example there were twelve elements, with subscripts from 1 to 12. In addition to the standard type of array, ATOM BASIC provides two other types of array called 'byte vectors' and 'word vectors'. Byte vectors are useful when only a small range of numbers are needed, and they use less storage space than word arrays. Word vectors use the same amount of storage as arrays, but can be manipulated in a more flexible manner.


7.1 Arrays - AA to ZZ The array in ATOM BASIC consists of a pair of identical letters a followed by the subscript in brackets: for example, EE(3). Each element in this type of array can contain numbers as large as the simple variables A to Z, namely, between about -2000 million and 2000 million. Before an array can be used space must be reserved for it by a DIM, or 'dimension', statement which tells BASIC how large the array - is to be. For example, to reserve space for an array called AA with the five elements AA(0), AA(1), AA(2), AA(3), and AA(4), the statement would be: DIM AA(4) The DIM statement allocates space for arrays starting at the first free memory location after the program text. If this were the first a DIM statement encountered in the program the element AA(0) would be at _45_
TOP, above the program text:
TOP:
? ? ? ?
^
AA(0)
^
AA(1)
^
AA(2)
^
AA(3)
????^
AA(0)
^
AA(1)
^
AA(2)
^
AA(3)
????
^
AA(0)
^
AA(1)
^
AA(2)
^
AA(3)
The question marks represent unspecified values, depending on what the 
array contained when it was dimensioned. If now another array were 
dimensioned with the statement:
DIM BB(3)
space for the array BB would be reserved immediately following on from 
AA.
Array elements can appear in expressions, and be assigned to, just 
like the simple variables A to Z. For example, to make the value of 
AA(3) become 776 we would execute:
AA(3)=776
Then we could execute:
AA(1)=AA(3)*2 
AA(0)=AA(3)-6
and so on. The resulting array would now be:
TOP:
770 1552 ? 776 ?
^
AA(0)
^
AA(1)
^
AA(2)
^
AA(3)
^
AA(4)

There are two places in BASIC programs where array elements may not be used; these are: l. As the control variable in a FOR...NEXT loop. 2. In an INPUT statement. In these two cases the simple variables, A to Z, must be used. 7.1.1 Histogram The following program illustrates the use of arrays to plot a histogram of the temperature over the twelve months of the year. The temperatures, assumed to be in the range 0 to 100, are first entered in and are stored in the array TT(1..12). 1 REM Histogram 10 DIM TT(12) 20 FOR 2=1 TO 12;INPUT K 30 TT(J)=K; NEXT J 40 PRINT $12; CLEAR 0; 0=5 50 MOVE 60,12; DRAW 12,12 60 DRAW 12,42 70 FOR N=11 TO 0 STEP -1 80 IF N=7 PRINT "TEMP." 90 IF N%2=0 PRINT N*10 100 PRINT';NEXT N 110 PRINT " JAN MAR MAY JUL SEP NOV"' 120 PRINT " FEB APR JUN AUG OCT DEC"' 130 PRINT " MONTH"' _46_
140 FOR N=1 TO 12; J=11+4*N 150 MOVE J,12; DRAW J,(TT(N)*3/10+12) 160 NEXT N; END Description of Program: 20-30 Input 12 values 40 Clear screen 50-60 Draw axes 70-100 Label vertical axis 110-130 Label horizontal axis 140-160 Plot histogram bars Program size: 415 bytes Array storage: 52 bytes


7.1.2 Sorting Program The following program illustrates the use of arrays to sort a series of numbers into ascending order. It uses a fairly efficient sorting procedure known as the 'Shell' sort. The program, as written, reads in 20 numbers, calls a subroutine to sort the numbers into order, and prints the sorted numbers out. 1 REM Sorting 5 DIM AA(20) 10 FOR N=1 TO 20; INPUT J 20 AA(N)=J; NEXT N 30 N=20; GOSUB s 40 FOR N=l TO 20; PRINT AA(N)' 50 NEXT N 60 END 100sM=N 110 DO M=(M+2)/3 120 FOR I=M+1 TO N 130 FOR J=I TO M+1 STEP -M 140 IF AA(J)>=AA(J-M) GOTO b 150 T=AA(J); AA(J)=AA(J-M); AA(J-M)=T 160 NEXT J 170b NEXT I 180 UNTIL M=l; RETURN Description of Program: 5-20 Read in array of numbers 30 Call Shell sort 40-50 Print out sorted array 100-180 s: Shell sort subroutine 140-150 Swap elements which are out of order. Variables: AA(1..20) - Array to hold numbers I,J - Loop counters N - Number of elements in array AA M - Subset step size T - Temporary variable Program size: 332 bytes Array storage: 84 bytes


7.1.3 Arbitrary-Precision Arithmetic The following program allows powers of two to be calculated to any precision, given enough memory. As it stands the program will calculate all the powers of 2 having less than 32 digits. The digits _47_
are stored in an array AA, one digit per array element. Every power of 2 is obtained from the previous one by multiplying every element in the array by 2, and propagating a carry when any element becomes more than one digit. 5 REM Powers of Two 10 DIM AA(31) 20 0=1; P=O 30 AA(0)=1 40 FOR J=1 TO 31 50 AA(J)=0 60 NEXT J 70 DO J=31 80 DO J=J-1; UNTIL AA(J)<>0 85 PRINT'"2"" P "=" 90 FOR K=J TO 0 STEP -1 94 PRINT AA(K) 96 NEXT K 110 C=O 120 FOR J=O TO 31 130 A=AA(J)*2+C 140 C=A/10 150 AA(J)=A%10 160 NEXT J 170 P=P+1 180 UNTIL AA(31)<>0 190 END Description of Program: 40-60 Zero array of digits 80 Ignore leading zeros 85-96 Print power 110-160 Multiply current number by 2 180 Stop when array overflows. Variables: AA - Array of digits; one digit per element C - Decimal carry from one digit to next J - Digit counter K - Digit counter P - Power being evaluated Program size: 356 bytes Array usage: 124 bytes Total memory: 480 bytes.


7.1.4 Digital Waveform Processing The following program uses a 256-element array to store a waveform which can be low-pass filtered, converted to a square wave, or printed out. 1 REM Digital Waveform Processing 5 DIM AA(255) 10 H=2000 15 CLEAR4 23 GOS.s; GOS.q 25 Z=160; GOS.p 28 GOS.1 30 Z=96; GOS.p 32 GOS.s 34 Z=32; GOS.p _48_
90 END 1000pREM Plot Waveform 1005 MOVE 0,96 1010 FOR N=O TO 255 1020 PLOT13,N,(Z+AA(N)/H) 1030 NEXT N 1040 RETURN 2000sREM Make Sine Wave 2010 S=O;C=40000 2020 FOR N=O TO 255 2030 AA(N)=-S 2040 C=C-S/10 2050 S=S+C/10 2060 NEXT N 2070 RETURN 3000qREM Make Square Wave 3010 FOR N=O TO 255 3020 IF AA(N)<=0 AA(N)=40000 3030 IF AA(N)<0 aa(n)="-40000 3035 NEXT N 3040 RETURN 4000lREM Low Pass Filter 4010" b="O 4020 FOR" n="O TO 255 4030" b="AA(N)*360/1000+B*697/1000 4040" aa(n)="B; NEXT N 4050 RETURN Description of Program: 23 Calculate a square wave 25 Plot it at top of screen 28 Low-pass filter the square wave 30 Plot it in centre of screen 32 Calculate a sine wave 34 Plot it at bottom of screen 1000-1040 p: Plots waveform 2000-2070 s: Calculates a sine wave. 3000-3040 q: Squares-up the waveform 4000-4050 l: Low-pass filters the waveform Variables: AA(0...255) - Array of points, values between -40000 and 40000. B - Previous value for low-pass filter C - Cosine of waveform H - Scalinq factor for plotting waveforms N - Counter S - Sine of waveform Z - Vertical coordinate for centre of waveform. Program size: 564 bytes. Array storage: 1024 bytes Total memory: 1588 bytes _49_

Sample plot: waveform.png


7.1.5 Subscript Checking Many BASIC interpreters perform extensive checking whenever an array element is used in a program. For example, if an array were dimensioned: DIM RR(10) then every time the array were used the subscript would be checked to make sure that it was both 0 or greater, and 10 or less. Obviously these two checks slow down the execution of a program, and so in ATOM BASIC only the first check is performed, so that only positive subscripts are allowed. It is left to the programmer to ensure that subscripts do not go out of range. Assigning to an array whose subscript is out of range will change the values of other arrays, or strings, dimensioned after that array. If required, the programmer can easily add array subscript checking; for example, if the array assignment were: RR(A)=35 the statement: IF A>10 THEN ERROR could be added before the assignment to cause an error if the array subscript, A, went out of range.


7.1.6 Multi-Dimensional Arrays The standard types of array in ATOM BASIC are one-dimensional. In other words, they have just one subscript, and so can be visualised as lying in a straight line; hence the name 'array'. Sometimes it is convenient to make each element of an array represent a cell in a square 'matrix'; each element would then have two subscripts corresponding to the column and row of that square. Such two-dimensional arrays are called 'matrices'. Consider the following representation of a 3 by 6 matrix: _50_

  0 1 2 3 4 5
0            
1            
2         x  

The whole matrix has 3 x 6 = 18 elements, and the element shown with an X would have the subscripts (2,4). ATOM BASIC does not have a direct representation for two-dimensional (or higher dimension) arrays, but they are easily represented using the single-dimension arrays AA to ZZ as described in the following sections.


7.1.7 Calculation of Subscripts To represent a two-dimensional matrix using a one-dimensional array imagine the matrix divided into rows as shown:
  0 1 2 3 4 5
0            
  0 1 2 3 4 5
1            
  0 1 2 3 4 5
2            

The first element of row 1, with subscripts (1,0), follows immediately after the last element of row 0, with coordinates (0,5). Consider the general case where the matrix has M rows numbered 0 to N-l, and N columns numbered 0 to N-1. The matrix can be dimensioned, using a one-dimensional array, with the DIM statement: DIM XX(M*N-1) Any array element, with subscripts A and B, can be referenced as: XX(A*N+B) In the earlier example the array had dimensions 3 x 6 and so would be dimensioned: DIM XX(17) The array element with subscripts (2,4) would be given by: xx(16)


7.1.8 Solving Simultaneous Equations The following program will solve a number of linear simultaneous equations, using a matrix to hold the coefficients of the equations, and a matrix inversion technique to find the solution. The program prints the solutions as integers, where possible, or as exact fractions. This method has the advantage over the standard pivotal condensation technique that for integer coefficients the answers are exact integers or fractions. The example run shown solves the pair of equations: a + 2b + 1 = 0 4a + 5b + 2 = 0 10 REM Simultaneous Equations 50 INPUT"NUMBER OF EQUATIONS="N 60 I=N*N;J=N*(N+1) _51_
65 DIM AA(I),CC(J),II(N) 70 0=0;FOR I=1TON;FOR J=1TO N+1 80 PRINT"C("I","J")=";INPUT C 90 CC((I-1)*(N+1)+J)=C;NEXT J;NEXT I 100 L=N+1;GOSUB c;E=D;M=l-2*(N%2) 110 PRINT'"SOLUTION:"' 112 IF E<0 e="-E;M=-M 115 IF" e="O;PRINTDEGENERATE!';END 120 FOR" l="1TON;GOSUB c 125" print"x("l")="130" a="M*D;B=E;DO" a="A%B 140 IF ABS(B)">ABS(A) THEN T=B;B=A;A=T 150 UNTIL B=O;A=ABS(A) 151 P.(M*D)/A;IF E/A<<1 PRINT"/"E/A 155 M=-M;PRINT';NEXT L;END 160cFOR I=1TON;FOR J=1TON;K=I*N-N+J 170 IF J<l aa(k)="CC(K+I-1) 180 IF J"<=L AA(K)=CC(K+I) 190 NEXT J;NEXT I 200dD=O;F=l;S=l 210 FOR J=1TON;II(J)=J;F=F*J;NEXT J 215 GOSUB f 220 FOR H=2TOF;GOSUB e;NEXT H;RETURN 230eI=N-1;J=N 240gIF II(I)>=II(I+1) I=I-1;GOTO g 250hIF II(I)>=II(J) J=J-1;GOTO h 260 GOSUB i;I=I+1;J=N;IF I=J GOTO f 270 DO GOSUB i;I=I+1;J=J-1;UNTIL I>=J 280fp=i;FOR K=1TON;P=P*AA(N*K-N+II(K)) 290 NEXT K;D=D+S*P;RETURN 300iK=II(I);II(I)=II(J);II(J)=K 310 S=-S;RETURN Description of Program: 50-60 Allocate space for matrix 70-90 Read in matrix of coefficients 120-155 Print solutions 130-150 Find GCD of solution, so it is printed in lowest terms 160-190 c: Permute terms to obtain next addition to determinant; i.e. for 5 equations, starting with (1,2,3,4,5) run through all permutations to (5,4,3,2,1). 280-290 f: Add in next product to determinant. 300-310 i: Swap terms in permutation. Variables: AA(1...N*N) - Matrix CC(1...N*N+N) - Matrix of coefficients S - Signature of permutation. Program Size: 932 bytes. Variable Space: (2*(N*N+N)+3)*4 bytes Sample run: <RUN NUMBER OF EQUATIONS=?2 C(1,1)=?1 C(1,2)=?2 C(1,3)=?1 C(2,1)=?4 C(2,2)=?5 C(2,3)=?2 _52_
SOLUTION: X(1)= 1/3 X(2)= -2/3


7.2 Byte Vectors Using, '?' It is sometimes wasteful of memory to allocate space for numbers over the range provided by word arrays so a second type of array representation is provided which only allocates one byte, rather than four bytes, for each array element. These are referred to as 'byte vectors', and they are in effect one-dimensional arrays. Byte vectors differ from word arrays in that they use one of the simple variables A to Z to hold the 'base' address of the array; i.e. the address in memory where the zeroth element of the array will reside. The array subscripts are simply 'offsets' from this base address; i.e. the subscript is added to the base address to give the address of the array element. The vector elements are written as: A?0, A?1, A?2, ... etc where A is the simple variable used to hold the base address of the vector, and the number following the question mark is the subscript. Note that the zeroth element of a byte vector, A?0, is equivalent to ?A, the contents of the location with address A. Similarly A?1 is equivalent to ?(A+1), and so on. Byte vectors can be dimensioned by the DIM statement; for example, to dimension a byte vector with elements from A?0 to A?11 the statement would be: DIM A(11) Because the DIM statement dimensions arrays and vectors from the end of the program onwards, the above DIM statement is equivalent to: T=ZOP; A=T; T=T+12 where T is a variable used to keep track of the next free memory location. Note that space for vectors can be reserved anywhere in memory, as distinct from arrays which can only be assigned from TOP onwards using the DIM statement. For example, to assign space for a vector S corresponding to the screen memory, simply execute: S=#8000 Elements of this vector would then correspond to locations on the screen; e.g. S?31 is the location corresponding to the top right-hand corner of the screen. Each element of a byte array can hold a positive number between 0 and 255, or a single character. Strings are simply byte vectors containing characters. Note that the subscript of a byte array can be an arbitrary expression provided that it is enclosed in brackets.


7.3 Word Vectors Using '!' A second representation for word arrays is provided in ATOM BASIC using the word indirection operator '!', and is mentioned here for completeness, although for simple problems involving arrays the word arrays AA to ZZ are probably more convenient. Word vectors are similar to the byte vectors already described, but each element of the vector consists of a word rather than a byte. Each element consists of the base address variable separated from the subscript, or offset, by a 'pling' '!'. Note that the subscript should be incremented by 4 for each element, since each element is offset 4 bytes from the previous one. For example, a word vector W might have the six elements: _53_
W!O, W!4, W!8, W!12, W!l6, W!20. Space can be dimensioned for word vectors by using the DIM statement, and allowing 4 bytes per element; for example, to provide storage for the above 6 elements, execute: DIM W(23) Note that the zeroth element of the vector, W!0, is equivalent to !W.


7.3.1 Prime Numbers The following program finds all the prime numbers up to 99999. It uses a word vector to store primes already found, and only tests new candidates for divisibility by these numbers: 1 REM Prime Numbers 10 @=8;S=4;Z=O;J=TOP;G=J;!G=3;PW+S 20 FORT=3T099999STEP2 30cIFT%!G=Z G=J;N. 40 IFT>!G*!G G=G+S;G.c 50 P.T;!P=T;G=J;P=P+S;N. 60 END Description of Program: 10 Set up vector 20 Test all odd numbers 30 If divisible, try another. 40 Have we tried enough divisors? 50 Must be prime - print it. Variables: !G - Divisor being tested J - Equal to TOP !P - Vector of divisors S - Bytes per word T - Candidate for prime Z - Constant zero. Program size: 155 bytes Vector: as required.


7.3.2 Call by Reference A major advantage of word vectors over the word arrays is that their base addresses are available as values, and so can be passed to subroutines. As an example, consider this program: 10 A=TOP; B=A+40 . . 90 P=A; GOSUB p; REM Output A 94 P=B; GOSUB p; REM Output B 98 END 100pREM Print 10 Elements of array P 105 @=8; PRINT I 110 FOR J=O TO 39 STEP 4 120 PRINT P!J 130 NEXT J 140 PRINT ' 150 RETURN _54_
In this example subroutine p can be used ta print any array by passing its base address over in the variable P; this is known as a 'call by reference' because the subroutine is given a reference to the array, rather than the actual values in the array.


7.3.3 Arbitrary Precision Powers The following program illustrates the use of word vectors to calculate the value of any number raised to any other number exactly, limited only by the amount of memory available. The program stores four decimal digits per word, so that the product of two words will not cause overflow, and the result is calculated as a word vector. 1 REM Arbitrary Precision Powers 5 T=#3BFF 10 H=(T-TOP)/3; DIM P(H),S(H),D(H) 15 H=10000 20 @=0;PRINT'" POWER PROGRAM" 30 PRINT'" COMPUTES Y X, WHERE x>0 AND Y>0" 40 INPUT'" VALUE OF Y"Y," VALUE OF X"X 50 IFX<10ry<1print” value out of range";run 60 n="Y;N=X;GOSUBp 70 PRINT Y" "x"="P!!P;IF!P<8 RUN 90" f.l="!P-4T04STEP-4 95 IFL!P<1000P.O 100 IFL!P<100P.O 110 IFL!P<10P.O 120" p.l!p;n.;run 140* 200pj="M;IFN%2=0J=1 210" r="P;GOS.e;J=M;R=S;GOS.e;IFN=1R. 250" b="S;DOA¿B;GOS.m;B=E 255" n="N/2;A=P;IFN%2GOS.m;P=E 260" u.n<2;r. 280* 300mld="IA+lB+4;F.J=4TOID+4S.4 310 D!J¿O;N.;W¿D-4 320 F.J¿4TO!B" s.4;c="O;G=BlJ 325" v="W+J;F.L¿4TO!A S.4 330" q¿a!l*g+c+v!l;v!l="q%h 340" c="Q/H;N.;V!LW;N. 370 DO!DE!0-4;U.D!!DC">0;E=D;D=A;R. 380* 400e!R=O;DO!R=lR+4;R!lR=J%H 410 JW/H;U.J<1;r. description of program: 5 set t to top of lower text space. 10 divide available memory between p, s, and d 20-40 read in values of y and x 50 disallow negative values 60 calculate power 70 print result if fits in one word 90 print rest of result, filling in leadinq zeros. 140 blank line to make listing clearer. 200-260 p: calculates power. looks at binary representation of x and for each bit squares b, and if bit is a 1 multiplies p by current b. 300-370 m: multiply together the vectors pointed to by a and b and put the result into the vector pointed to by d. pointers to vectors get changed; e points to result. _55_
400-410 e: Unpack J into vector pointed to by R; store number of words in !R. Variables: D!0... - Workspace vector H - Radix for arithmetic P!1... - Vector for unpacked result !P - Number of elements used in P S!0... - Workspace vector T - Top of available memory Program size: 733 bytes. Additional storage: as available. Sample run: >RUN POWER PROGRAM COMPUTES Y"X, WHERE X>0 AND Y>0 VALUE OF Y?16 VALUE OF X?64 16^64=1157920892373161954235709850086879078532699846656405640394575 84007913129639936


7.3.4 Vectors of Vectors A second way of representing two-dimensional arrays is possible using the ATOM's indirection operators '?' and '!'; this avoids the need for a multiplication to calculate the subscript, but does require slightly more storage. The idea is to think of a two-dimensional matrix as a vector of vectors; first a vector is created containing the addresses of the rows of the matrix. For example, for a matrix called X with columns 0 to M, and rows 0 to N, the following statements will set up the vector of row addresses: DIM X(2*N-1) FOR J=O TO N*2 STEP 2; DIM Q(m); X!J=Q; NEXT J A word array is used to hold the base addresses. Q is a variable used to hold the base address temporarily. Now that the vector of row base addresses has been set up, the element with subscripts A,B is: X!(A*2)?B _56_

8 strings

A 'string' is a sequence of characters; the characters can be anything letters, digits, or punctuation marks. They can even be control characters.


8.1 Quoted Strings Strings are represented in a program by enclosinq the characters between quotation marks; quoted strings have already been introduced in the context of the PRINT and INPUT statements. For example: "THIS IS A STRING" To represent a quotation mark in a quoted string the quotation mark is typed twice. Valid strings always contain an even number of quotation marks. For example: PRINT"HE SAID: ""THIS IS A VALID STRING""" will print: HE SAID: "THIS IS A VALID STRING"


8.2 String Variables The variables A to Z have already been met, where they are used to represent numbers. These variables can also be used to represent strings, and strings can be manipulated, input with the INPUT statement, printed with the PRINT statement, and there are several functions for manipulating strings.


8.2.1 Allocating Space for Strings BASIC allows strings of any size up to 255 characters. To use string variables space for the strings should first be allocated by means of a DIM (dimension) statement. For example, for a string of up to 10 characters using the variable A the statement would be: DIM A(10) Any number of strings can be dimensioned in one DIM statement.


8.2.2 String Operator '$' Having allocated space for the string it can then be assigned a value. For example: $A="A STRING" The '$' is the string-address operator. It specifies that the value following it is the address of the first character of a string. The effect of the statement DIM A(10) is to reserve 11 memory locations in the area of free memory above the text of the BASIC program, and to put the address of the first of those locations into A. In other words, A is a pointer to that area of memory. After the above assignment the contents of those locations are as follows: _57_

A:
A   S T R I N G ~ ? ?

The question-marks indicate that the last two locations could contain anything. The character ' ' represents 'return' which is automatically stored in memory to indicate the end of the string. The DIM statement allocates one extra location to hold this terminator character, although you will not normally be aware of its presence. Note that it would be dangerous to allocate a string of more than 10 characters to A since it would exceed the space allocated to A.


8.2.3 Printing strings A string variable can be printed by writing: PRINT $A This would print: A STRING> and no extra spaces are inserted before or after the string.


8.2.4 String Assignment Suppose that a second string is dimensioned as follows: DIM B(8) The string $A can be assigned to $B by the statement: $B=SA which should be read as 'string B becomes string A'. The result of this assignment in memory is as follows:
A:
A   S T R I N G ~ ? ? A   S T R I N G ~
^
A
                    ^
B
               


8.2.5 String Equality It is possible to test whether two strings are equal with the IF statement. For example: $A="CAT"; $B="CAT" IF $A=$B PRINT "SAME" would print SAME.


8.2.6 Strinq Input The INPUT statement may specify a string variable, in which case the string typed after the '?' prompt, and up to the ’return', will be assigned to the string variable. The maximum length of line that can be typed in to an INPUT statement is 64 characters so, for safety, the string variable in the INPUT statement should be dimensioned with a length of 64.


8.3 String Functions Several functions are provided to help with the mainpulation of strings. _58_


8.3.1 Length of a String - LEN The LEN function will return the number of characters in the string specified in its argument. For example: $A="A STRING" PRINT LEN(A) will print the value 8. Note that: $B="""" PRINT LEN(B) will print 1 since the string B contains only a single quote character.


8.3.2 CH The CH function will return the ASCII value of the first character in the string specified by its argument. Thus: CH"A" will be equal to 65, the ASCII code for A. The string terminating character 'return' has a value of 13, so: CH"" will be equal to 13.


8.4 String Manipulations The following sections show how the characters within strings can be manipulated, and how strings can be concatenated into longer strings or broken down into substrings.


8.4.1 Character Extraction - '?' Individual Characters in a string can be accessed with the question-mark ’?' operator. Consider again the representation of the string A. Number the characters, starting with zero:
 
A   S T R I N G ~ ? ?
0
^
A
1 2 3 4 5 6 7 8 9 10
The value of the Nth. character in the string is then simply A?N. For example, A?7 is "G", etc. In general A?B is the value of the character stored in the location whose address is A+B; therefore A?B is identical to B?A. In other words, a string is being thought of as a byte vector whose elements contain characters; see section 7.2. The following program illustrates the use of the '?’ operator to invert all the characters in a string which is typed in: 1 REM Invert String 5 DIM Q(64) 10 INPUT $Q 20 FOR N=O TO LEN(Q)-1 30 Q?n=q?n fg 020 40 NEXT N 50 PRINT $Q 60 RUN _59_


8.4.2 Encoding/Decoding Program As a slightly more advanced example of string operations using the ’?' operator, the following program will produce a very secure encoding of a message. The program is given a number, which is used to 'seed' BASIC's random number generator. To decode the text the negative of the same seed must be entered. 1 REM Encoder/Decoder 10 S=TOP; ?12=0 20 INPUT'"CODE NUMBER"T 30 !8=ABS(T) 40 INPUT'$S 50 FOR P=S TO S+LEN(S) 60 IF ?P<041 goto 100 70 r="ABS(RND)%26 80 IF T<0 THEN" r="26-R 90" ?p="(?P-041+R)%26+#41 100 NEXT P 110 PRINT $S 120 GOTO 40 Description of Program: 20 Input code number 30 Use code nurrber to seed random number generator 40 Read in line of text 50-100 For each cl aracter, if it is a letter add the next random number to it, modulo 26. 110 Print out encoded string. Variables: P - Address of character in string R - Next random number S - Address of string; set to TOP. T - Code number Program size: String storage: up to 64 bytes Sample run:">RUN CODE NUMBER?123 ?MEETING IN LONDON ON THURSDAY BGYKPYI CM NHSHVO VU RGFGDHJI <RUN CODE NUMBER?123 ? MEETING IN LONDON ON THURSDAY BGYKPYI CM NHSHVO VU RGFGDHJI ? < <RUN To illustrate how secure this encoding algorithm is you may like to attempt to find the correct decoding of the following quotation: YUVHW ZY WKQN IAVUAG Qm SHXTSDK GSY IEJB RZTNOL UFQ FTONB JB BY CXRK Qcjf UN TJRB. SWB FJA IYT WCC LQFWHA YHW OHRMNI OUJ _60_
HTJ I TYCU GQYFT FT SGGHH HJ FRP ELPHQMD, RW LN Qohd OQXSER CUAB. DKLCLDBCV.


8.4.3 Concatenation Concatenation is the operation of joining two strings together to make one string. To concatenate string B to the end of string A execute: $A+LEN(A)=$B For example: 10 DIM A(10),B(5) 20 $A="ATOM" 30 $B="BASIC" 40 $A+LEN(A)=$B 50 PRINT $A 60 END will print: ATOMBASIC>


8.4.4 Right-String Extraction The right-hand part of a string A, starting at character N, is simply: $A+N For example, executing: 10 DIM A(10),B(5) 20 $A="ATOMBASIC" 30 $B=$A+4 40 END will give string B the value "BASIC".


8.4.5 Left-String Extraction A string A can be shortened to the first N characters by executing: $A+N="" Since the 'return' character has the value 13, this is equivalent to: A?N=13


8.4.6 Mid-String Extraction The middle section of a string can be extracted by combining the techniques of the previous two sections. For example, the string consisting of characters M to N of string A is obtained by: $A+N=""; $A=$A+M For example:, if the following is executed: 10 DIM A(10) 20 $A="ATOMBASIC" 30 $A+5=0"; $A=SA+1 40 END then string A will have the value "TOMB". _61_


8.5 Arrays of Fixed-Length Strings The arrays AA to ZZ may be used as string variables, thus providing the ability to have arrays of strings. To allocate space for an array of strings the DIM statement can be incorporated into a FOR...NEXT loop. For example, the followinq program allocates space for 21 strings, AA(0) to AA(20), each capable of holding 10 characters: 25 DIM AA(20) 35 FOR N=O TO 20 40 DIM J(10) 50 AA(N)=J 60 NEXT N Note the use of a dummy variable J to allocate the space for each string. Individual elements of the string array can then be assigned to as follows: $AA(0)="ZERO" $AA(10)="TEN" and so on.


8.5.1 Day of Week The following program calculates the day of the week for any date in the 20th. century. It stores the names of the days of the week in a string array. 1 REM Day of Week 10 DIM AA(6) 20 FOR N=O TO 6; DIM B(10); AA(N)=B; NEXT N 30 $AA(0)="SUNDAY”; $AA(1)=”MONDAY" 40 Saa(2)="tuesday";$aa(3)="wednesday" 50 $AA(4)="THURSDAY";$AA(5)="FRIDAY" 60 $AA(6)="SATURDAY" 70 INPUT"DAY OF WEEK"'’"YEAR "Y,"MONTH ”M,"DATE IN MONTH "D 80 Y=Y-1900 90 IF Y<0 or y<99 PRINT"ONLY 20TH CENTURY I"';GOTO 70 100 IF M>2 THEN M=M-2; GOTO 120 110 Y=Y-1; M=M+10 120 E=(26¿M-2)/10+D+Y+Y/4+19/4-2*19 130 PRINT"IT IS "$AA(ABS(E%7)) 140 END Description of Program: 10-20 Allocate space for string array 30-60 Set array elements 70 Input date 80-120 Calculate day 130 Print day of week. Variables: $AA(0...6) - String array to hold names of days B - Temporary variable to hold base address of each string D - Date in month E - Expression which, modulo 7, gives day of week. M - Month N - Counter Y - Year in 20th. century. Program size: 458 bytes. Array storage: 105 bytes. _62_
Total memory: 563 bytes.


8.6 Arrays of Variable-Length Strings The most economical way to use the memory available is to allocate only as much space as is needed for each string. For example the following program reads in 10 strings and saves them in strings called VV(1) to VV(10): 10 DIM VV(10),T(-1) 20 FOR N=1 TO 10 30 INPUT $T 40 VV(N)=T 50 T=T+LEN(T)+1 60 NEXT N 70 INPUT "STRING NUMBER",N 80 PRINT $VV(N),' 90 GOTO 70 The statement DIM T{-1) sets T to the address of the first free memory location. T is then incremented past each string to the next free memory location as each string is read in. Finally, when 10 strings have been read in the program prompts for a string number and types out the string of that number. For example, if the first three strings entered were: "ONE", "TWO", and "THREE”, the contents of memory would be:
 
O N E ~ T W O ~ T H R E E ~ ? ? ?
^
VV(1)
^
VV(2)
^
VV3
^
T


8.7 Reading Text Some BASICs have statements READ and DATA whereby strings listed in the DATA statements can be read into a string variable using the READ statement. Although ATOM BASIC does not provide these actual statements, reading strings specified as text is a fairly simple matter. The following program reads the strings "ONE", "TWO" ... etc. into a string variable, $A, and prints them out. The strings for the numbers are specified as text after the program. They are identified by a label 't', and a call to subroutine 'f' sets Q to the address of the first string. Subroutine 'r' will then read the next string from the list: 10 REM Read Text 20 DIM A(40); L=CH"t" 25 GOSUB f 30 FOR J=l TO 20; GOSUB r 40 PRINT $A 50 NEXT J 60 END 500fREM point Q to text 510 Q=218*256 520 DO Q=Q+1 530 UNTIL ?Q=0d AND Q?3=l 540 Q=Q+4; RETURN 550* 600rREM read next entry into A _63_
605 REM changes: A,Q,R 610 R=-1 620 DO R=R+1; A?R=Q?R 630 UNTIL A?R=CH"," OR A?R=0D 640 IF A?R=fD Q=Q+3 650 Q=Q+R+1; A?R=0D; RETURN 660* 800tONE,TWO,THREE,FOUR,FIVE 810 SIX,SEVEN,EIGHT,NINE,TEN 820 ELEVEN,TWELVE,THIRTEEN 830 FOURTEEN,FIFTEEN,SIXTEEN 840 SEVENTEEN,EIGHTEEN,NINETEEN 850 TWENTY Description of Program: 25 Find the text 30 Read in the next string 40 Print it out 500-550 f.: Search for label t and point Q to first string 600-660 r: Read up to comma or return and put string into $A 800-850 t: List of 20 strings Variables: $A - String J - Counter L - Label for text Q - Pointer to strings R - Temporary pointer Program size: 511 bytes String storage: 41 bytes Total memory: 552 bytes. The program can be moaified to read from several different blocks of text with different labels by changing the value of L. Also note that the character delimiting the strings may be any character, specified in the CH function in line 630.


8.7.1 Reading Numeric Data Numeric data can be specified as strings of characters as in in the Read Text program of the previous section, and converted to numbers using the VAL command in the extension ROM. For example, modify the Read Text program by changing line 40 to: 40 FPRINT VAL A and provide numeric data at the label 't', for example as follows: 800t1,2,3,4,1E30,27,66 810 91,1.2,1.3,1.4,1.5 820 13,14,15,16,17 830 18,19,20


8.8 Printing Single Characters - '$' A special use of the '$' operator in the PRINT statement is to print characters that can not conveniently be specified as a string in the program, such as control characters and graphics symbols. Normally '$’ is followed by a variable used as the base address of the string. If, however, the value following the dollar is less than 255, the character corresponding to that code will be printed instead. The following table gives the control codes, characters, and graphics symbols corresponding to the different codes: _64_
Hex: Decimal: Character Printed: #00 - #1F 0 - 31 Control codes #20 - 05F 32 - 95 ASCII cnaracters #60 - #9F 96 - 159 Inverted ASCII characters #A0 - #DF 160 - 223 Grey graphics symbols 4EO - #FF 224 - 255 White graphics symbols Note that only half of the 64 possible white graphics symbols can be obtained in this way. The most useful control codes are specified in the following sections; for a full list of control codes see section 18.1.3.


8.8.1 Cursor Movement The cursor can be moved in any of the four directions on the screen using the following codes: Hex: Decimal: Cursor Movement: #08 8 Left #09 9 Right #OA 10 Down #Os 11 Up The screen is scrolled when the cursor is moved off the bottom line of the screen; the cursor cannot be moved off the top of the screen. Note that the entire screen memory is modified by scrolling; every line is shifted up one line, and the bottom line is filled with spaces.


8.8.2 Screen Control The following control codes are useful for controlling the VDU screen: Hex: Decimal: Control Character: #OC 12 Clear screen and home cursor #1E 30 Home cursor to top left of screen


8.8.3 Random Walk The following program prints characters on the screen following a random walk. One of the cursor control codes, chosen at random, is printed to move the cursor; a white graphics character, chosen at random, is then printed followed by a backspace to move the cursor back to the character position. 1 REM Random Walk 10 DO 20 PRINT $ABS(RND)%4+8, $(0AO+ABS(RND)1#40), $8 30 UNTIL 0 _65_
_66_

9 Reading and Writing Data

The reader should now be familiar with the three types of data that can be manipulated using ATOM BASIC, namely: 1. Words i.e. numbers between -2000 million and 2000 million (approximately). Storage required: 4 bytes e.g. variables A to Z arrays AA(1) ... etc. word vectors A!4 ...etc. indirection !A ...etc. 2. Bytes i.e. numbers between 0 and 255, or single characters, or logical values. Storage required: 1 byte e.g. byte vectors A?1 ... etc. indirection ?A ...etc. 3. Strings i.e. sequences of between 0 and 255 characters, followed by a 'return'. Storage required: Length+1 bytes e.q. quoted string "A STRING" string variable $A ...etc. All these types of data can be written to cassette and read from cassette, making it very simple to make files of data generated by programs. The ATOM BASIC functions and statements for cassette input and output are designed to be fully compatible with the disk operating system, should that be added at a later stage. When the disk operating system is used, several files can be used by one program, and the individual files are identified by a 'file handle', a number specifying which file is being referred to. Although this facility is not available when working with a cassette system, the file handle is still required for compatibility.


9.2 Output To output a word to cassette the PUT statement is used. Its form is: PUT A,W where A and W are the file handle, and word for output, respectively. To output a byte to cassette the BPUT statement is used; the form 1S: BPUT A,B where A is the file handle, and B is the byte for output. To output a string the SPUT statement is used. The form is: SPUT A,S where A is the file handle, and S is the base address of the string. _67_


9.3 Input To read a word from cassette the GET function is used. Its form is: GET A where A is the file handle. The function returns the value of the word. To read a byte the BGET function is used. Its form is: BGET A where A is the file handle. The BGET function returns the value of the byte, and can therefore be used in expressions; for example: PRINT BGET A + BGET A will read two bytes from cassette and print their sum. To read strings the SGET statement is used. The form is: SGET A, S where A is the file handle, and S is the base address where the string will be stored. The string S should be large enough to accomodate the string being read. Note the difference between SGET, which is a statement, and the functions BGET and GET; SGET cannot be used in expressions.


9.4 Find Input and Find Output The functions FIN (find input) and FOUT (find output) can optionally be called before inputting from, or outputting to, cassette. The functions are called with a null string as the argument, and they return the value 13; when used with a disk system the argument is the file name, and the value returned is the file handle. The FOUT function is called as follows: A=FOUT"" and it will cause the message: RECORD TAPE to be printed, and the program will wait for a key to be pressed before continuing execution. The FIN function is called as follows: A=FIN"" and it causes the message: PLAY TAPE to be printed, and again the program will wait for a key to be pressed. A dummy variable, such as A in this example, should be used to hold the file handle.


9.4.1 Data on Cassette The following program prompts for a series of values, terminated by a zero, and saves them on a cassette tape. The first byte saved on the tape is the number of words of data saved. 1 REM Data to Cassette 10 DIM VV(20) 20 N=O 30 DO INPUT J 40 VV(N)=J; N=N+1 50 UNTIL J=O OR N>20 60 A=FOUT"" _68_
70 BPUT A,(N-1) 80 FOR M=O TO N-1 90 PUT A,VV(M) 100 NEXT M 110 END Description of Program: 30-50 Input numbers 60 Warn user to start tape 70 Output number of bytes 80-100 Save values on cassette Variables: A - Dummy file handle J - Temporary variable for values input M - Counter N - Counter for number of values VV(0...20) - Array of numbers The next program reads the values back in and plots a histogram of the values. The program automatically scales the values if they are too large to fit onto the screen. 1 REM Plot Histogram from Cassette 10 DIM VV(20) 20 A=FIN""; N=BGET A 30 FOR M=O TO N 4O VV(M)= GET A 50 NEXT M 60 REM X=Maximum, Y=Minimum 70 X=VV(0); Y=VV(0) 80 FOR M=1 TO N 90 IF X<vv(k) then x="VV(M) 100 IF Y"<VV(M) THEN Y-VV(M) 110 NEXT M 120 S=(X-Y+63)/64 130 REM Plot Histogram 135 CLEAR 0 140 FOR M=O TO N 150 MOVE O,M 160 DRAW ((VV(M)-Y)/S),M 170 NEXT M 180 GOTO 180 Description of Program: 20-50 Read values into array 70-110 Find maximum and minimum values in array 120 Calculate scaling factor 140-170 Plot scaled histogram 180 Wait for ESC key. Variables: A - Dummy file handle M - Counter N - Number of values in array S - Scale factor for array VV(0...20) - Array of values X - Maximum value Y - Minimum value _69_


9.5 Reading and Writing Speed When writing data to the cassette it is important to remember that the program reading the data back will not be able to control the cassette; it will have to read the data before it has passed under the tape head. If the program to read the data will spend a substantial time between reading, it may miss bytes passing under the tape head unless a delay is inserted between bytes when writing to tape. As a general guide, the program to read the data should take no longer to read each byte than the program to write the data takes to write it.


9.6 Animal Learning Program The following program illustrates how a computer can be 'taught' information, so that a 'database' of replies to questions can be built up. The computer plays a game called 'Animals'; the human player thinks of an animal and the computer tries to guess it by asking questions to which the answer is either 'yes' or 'no'. Initially the computer only knows about a dog and a crow, but as the game is played the computer is taught about all the animals that it fails to guess. The program uses the cassette input/output statements to load the database, or tree, from cassette at the start of the game, and to save the enlarged database at the end of the game. First create a database by typing: GOSUB 9000; and record the database on a cassette. Then RUN the program and load the database you have just recorded. When the reply ’NO’ is given to the question 'ARE YOU THINKING OF AN ANIMAL' the program will save the new, enlarged, database on cassette. Also given is a sample run which was obtained after several new animals had been introduced to the computer. 1 REM Animals 10 REM Load Tree 20 F=FIN"" 23 DO UNTIL BGET F=#AA 25 FOR T=TOP TO TOP+GET F 30 ?T=BGET F; NEXT T 35 DO X=TOP 40 PRINT'"ARE YOU THINKING OF AN ANIMAL" 45 GOSUB q 48 IF Q=O THEN GOSUB z; END 50 DO PRINT $X+1 60 GOSUB q 65 P=X+LENX+1+Q; X=!P+TOP 70 UNTIL ?X<CH"*" 75 PRINT"IS IT "$X 80 GOSUB q 85 IF Q=4 PRINT "HO-HO";UNTIL 0 90 DO INPUT"WHAT WERE YOU THINKING OF"ST 95 UNTIL LEN T>2 98 L=T; GOSUB s 100 PRINT" TELL ME A Question" 110 PRINT"THAT WILL"'"DISTINGUISH" 120 PRINT "BETWEEN "Sl" AND " Sx ' 130 $T="*"; R=T+1 140 INPUT $R; !P=T-TOP; GOSUB s 145 K=T; T=T+8; GOSUB j 150 GOSUB q _70_
160 K!Q=X-TOP; K!(4-Q)=L-TOP 170 UNTIL 0 1000qINPUT $T 1010 IF ?T=CH"Y"THEN Q=4; RETURN 1020 IF ?T=CH"Q"THEN END 1030 Q=O; RETURN 2000j$T=$R; A=i 2010 DO A=A+1 2020 V=T?(A+4); ST+A+4="" 2030 IF $T+A=" IT "UNTIL 1; GOTO k 2035 T?(A+4)=V 2040 UNTIL A=LEN T-5 2100 PRINT"WHAT WOULD THE ANSWER BE"' 2110 PRINT"FOR " $X 2120 RETURN 2150kT?(A+4)=V; $T+A+1="" 2160 PRINT $T,$X,$T+A+3 2170 RETURN 3000sT=T+LEN T+1; RETURN 9000 REM Set-Up File 9010 T=TOP; ST="*DOES IT HAVE FOUR LEGS" 9015 GOSUB s; P=T; T=T+8; !P=T-TOP 9020 $T="A CROW"; GOSUB s; P!4=T-TOP 9025 $T="A DOG"; GOSUB s 9100zREN Save Tree 9110 F=FOUT "1' 9112 BPUT F,OAA; WAIT 9115 PUT F,(T-TOP-1) 9120 FOR N=TOP TO T-1 9130 RPUT F, ?N 9140 NEXT N 9150 RETURN Description of Program: 20-30 Load previous tree 23 Look for start flag 35 Reset X to top of tree 50 Print next question 70 Carry on until not a question 75 Guess animal 90-95 Wait for a sensible reply 98 Find end of reply 1000-1030 q: Look for Y, N, or Q; set Q accordingly 2000-2120 j: Look for "IT "in question and print question with "IT" replaced by name of animal. 3000 s: Move T to end of string $T. 9000 Set up tree file 9100 z: Save tree file. Variables: F - Dummy file handle K - Pointer to addresses of next two branches of tree L - Pointer to animal typed in P - Pointer to address of next question or animal. Q - Value of reply to question; no=0, yes=4. R - Pointer to question typed in T - Pointer to next free location X - Pointer to current position on tree Program size: 1254 bytes Additional storage: as required for tree. _71_
Sample run: >RUN ARE YOU THINKING OF AN ANIMAL?Y DOES IT HAVE FOUR LEGS?Y CAN YOU RIDE IT?N DOES IT HAVE STRIPES?N IS IT A DOG?N WHAT WERE YOU THINKING OF?A MOUSE TELL NE A Question THAT WILL DISTINGUISH BETWEEN A MOUSE AND A DOG ?DOES IT SQUEAK DOES A DOG SQUEAK?NO ARE YOU THINKING OF AN ANIMAL?Y DOES IT HAVE FOUR LEGS?Y CAN YOU RIDE IT?N DOES IT HAVE STRIPES?N DOES IT SQUEAK?Y IS IT A MOUSE?Y HO-HO ARE YOU THINKING OF AN ANIMAL?N RECORD TAPE > _72_

10 More Space and More Speed

This chapter shows how to abbreviate programs so that they will fit into a smaller amount of memory, and how to write programs so that they will run as fast as possible.


10.1 Abbreviating BASIC Programs Most versions of BASIC demand a large amount of redundancy. For example, the command PRINT must usually be specified in full, even though there are no other statements beginning with PR. In ATOM BASIC it is possible to shorten many of the statement and function names, and omit many unnecessary parts of the syntax, in order to save memory and increase execution speed. The examples in this manual have avoided such abbreviations because they make the resulting program harder to read and understand, but a saving of up to 30% in memory space can be obtained by abbreviating programs as described in the following sections.


10.1.1 Statements and Functions All statement and function names can be abbreviated to the shortest sequence of characters needed to distinguish the name, followed by a full stop. The following abbreviations are possible: Name: Abbreviation: ABS A. AND A. BGET B. BPUT B. CH CLEAR COUNT C. DIM DO DRAW END E. EXT E. FIN F. FOR F. FOUT FO. GET G. GOSUB GOS. GOTO G. IF INPUT IN. LEN L. LET L. LINK LI. LIST L. LOAD LO. MOVE NEW N. NEXT N. _73_
OLD OR PLOT PRINT P. PTR PUT REM RETURN R, RND R. RUN SAVE SA. SGET S. SHUT SH. SPUT SP. STEP S. THEN T. TO TOP T. UNTIL U. WAIT


10.1.2 Spaces Spaces are largely irrelevant to the operation of the BASIC interpreter, and they are ignored when encountered in a program. Their only effect is to cause a 13 microsecond delay in execution. There is "one place where a space is necessary to avoid an ambiguity as in the following example: FOR A=B TO C where the space after B is compulsory to make it clear that B is not the first letter of a function name.


10.1.3 LET Some BASICs demand that every assignment statement begin with the word LET; e.q.: LET A=B In ATOM BASIC the LET statement may be omitted, with a decrease in execution time.


10.1.4 THEN The word THEN in the second part of an IF statement may be omitted. For example: IF A=B C=D is perfectly legal. However, note that if the second statement begins with a T, or a '?' or '!' unary operator, some delimiter is necessary: IF A=B THEN T=Q Alternatively a statement delimiter ';' can be used as the delimiter: IF A=B; T=Q


10.1.5 Brackets Brackets enclosing a function argument, or an array identifier, are unnecessary and may be omitted when the argument, or array subscript, is a single variable or constant. For example, AA(3) may be written AA3, ABS(RND) may be written ABSRND, but AA(B+2) cannot be abbreviated. _74_


10.1.6 Commas The commas separating elements in a PRINT statement can be omitted when there is no ambiguity. For example: PRINT A,B,C,"RESULT",J may be shortened to: PRINTA B C"RESULT"J Note that the comma in: PRINT &A,&B is, however, necessary to distinguish the numbers from the single number (A&B) printed in hex.


10.1.7 Multi-Statement Lines Each text line uses one byte per character on the line, plus two bytes for the line number and a one-byte terminator character; thus writing several statements on one line saves two bytes per statement. Note that there are two occasions where this cannot be done: l. After an IF statement, because the statements on the line following the IF statement would be skipped if the condition turned out false. 2. Where the line number is referred to in a GOTO or GOSUB statement.


10.1.8 Control Variable in NEXT The FOR...NEXT control variable may be omitted from the NEXT statement; the control variable will be assumed to be the one specified in the most recently activated FOR statement.


10.2 Maximising Execution Speed ATOM BASIC is one of the fastest BASIC interpreters available, and all of its facilities have been carefully optimised for speed so that calculations will be performed as quickly as possible, and so that real-time graphics programs are feasible. To obtain the best possible speed from a program the following hints should be borne in mind; but note that many of these suqgestions reduce the legibility of the program, and so should only be used where speed is critical. l. Use the FOR...NEXT loop in preference to an IF statement and a GOTO. 2. Use labels, rather than line numbers, in GOTO and GOSUB statements. 3. Avoid the use of constants specified in the body of programs; instead use variables which have been set to the correct value at the start of the program. For example, replace: A=A*1000 by: T=1000 . . A=A*T 4. Write statements in-line, rather than in subroutines, when the subroutines are only called once, or when the subroutine is shorter than two or three lines. _75_
5. If a calculation is performed every time around a loop, make sure that the constant part of the calculation is performed only once outside the loop. For example: FOR J=1 TO 10 FOR K=1 TO 10 VV(K)=VV(J)*2+K NEXT K NEXT J could be written as: FOR J=1 TO 10 Q=vv(j)*2 FOR K=1 TO 10 VV(K)=Q+K NEXT K NEXT J 6. Where several nested FOR...NEXT loops are being executed, and the order in which they are performed is not important, arrange them so that the one executed the greatest number of times is at the centre. For example: FOR J=1 TO 2 FOR K=1 TO 1000 . . NEXT K NEXT J is faster than: FOR K=1 TO 1000 FOR J=1 TO 2 . . NEXT J NEXT K because in the second case the overhead for setting up the inner loop is performed 1000 times, whereas in the first example it is only performed twice. 7. Choose the FOR...NEXT loop parameters so as to minimise calculations inside the loop. For example: FOR N=O TO 9 DRAW AA(2*N), AA(2*N+1) NEXT N could be rewritten as the faster: FOR N=O TO 18 STEP 2 DRAW AA(N),AA(N+1) NEXT N 8. Use word operations rather than byte operations where possible. For example, to clear the graphics screen to white it is faster to execute: _76_
FOR N=#8000 TO #9800 STEP 4; !N=-1; NEXT N than the following: FOR N=08000 TO #9800; ?N=-1; NEXT N 9. The IF statement containing several conditions linked by the AND connective, as, for example: IF A=2 AND B=2 AND C=2 THEN ..... will evaluate all the conditions even when the earlier ones are false. Rewriting the statement as: IF A=2 IF B=2 IF C=2 THEN ..... avoids this, and so gives faster execution. _77_
_78_

11 Advanced Graphics

The ATOM provides nine different graphics modes, up to a resolution of 256x192 in black and white, and 128x192 in four selectable colours. The graphics modes use the BASIC statements PLOT, DRAW, and MOVE in an identical way. All the black-and-white graphics commands are present in the unexpanded ATOM, although extra memory will be required for the higher-resolution graphics modes. Colour plotting requires the addition of an assembler routine, or the COLOUR statement provided in the extension ROM.


11.1 Graphics Modes The nine graphics modes are listed below: Mode: Resolution: Memory: X: Y: 0 64 48 0.5 K 1a 64 64 1 K 1 128 64 1 K 2a 128 64 2 K 2 128 96 1.5 K 3a 128 96 3 K 3 128 192 3 K 4a 128 192 6 K 4 256 192 6 K


11.2 CLEAR

This statement clears the screen and puts it into graphics mode. It is followed by a number, or expression in brackets, to specify the mode. The graphics screen is labelled as follows:

^
|
Y
|
atom_1.png
0,0 -- X -->

The smallest square which can be plotted on the display is referred to as a 'pixel' (or 'picture element').


11.3 PLOT The graphics statements include a versatile 'PLOT K,X,Y' statement, the value of K determining whether to draw or move, plot lines or points, whether to set, clear, or invert, and whether to take the parameters X and Y as the absolute screen position, or as a displacement from the last point. The values K, X, and Y can be arbitrarily-complicated expressions. _79_
K: Function: 0 Move relative to last position 1 Draw line in white relative to last position 2 Invert line relative to last position 3 Draw line in black relative to last position 4 Move to absolute position 5 Draw line in white to absolute position 6 Invert line to absolute position 7 Draw line in black to absolute position 8 Move relative to last position 9 Plot point in white relative to last position 10 Invert point relative to last position 11 Plot point in black relative to last position 12 Move to absolute position 13 Plot point in white at absolute position 14 Invert point at absolute position 15 Plot point in black at absolute position


11.4 DRAW and PDVE In addition DRAW and MOVE statements are provided as convenient aliases for drawing a line and moving to an absolute X,Y position. MOVE X,Y is equivalent to PLOT 12, X, Y. DRAW X,Y is equivalent to PLOT 5, X, Y.


11.4.1 Random Rectangles The following program illustrates the use of relative plotting using the PLOT statement, and draws random rectangles on the display. The program will work in any of the graphics modes. 10 REM Random Rectangles 13 S=20 16 Z=1;B=O 17 W=64;H=48 18 E=W-S;F=H-S 20 CLEARB 30 FORQ=OT07 32 MOVE(ABSRND%E),(ABSRND%F) 35 C=ABSRND%S+1;D=ABSRNDSS+1;GOSUBs 37 NEXTQ;FOR Q=ot020000;nextq 38 GOT020 100sPLOTZ,C,O 110 PLOTZ, 0, D 120 PLOTZf CfO 130 PLOTZ,O,-D 140 RETURN Description of Program: 13-18 Set up constants 20 Initialise graphics 30 Draw 41 rectangles 32 Move to random point, leaving margin for size of largest rectangle. 35 Choose random rectangle 37 Wait; then repeat. 100 140 s: Draw rectangle. _8O_
Variables: C,D - Dimensions of rectangle E,F - Dimensions of safe part of screen to start drawing rectangle. H - Screen height Q - Counter S - Size of squares W - Screen width Z - Plot mode; draw relative. Program size: 278 bytes


11.5 Advanced Graphics Examples The following examples are designed for use with the higher-resolution graphics modes, and illustrate some of the applications that are possible using the ATOM's graphics facilities.


11.5.1 The Sierpinski Curve This curve is of interest to mathematicians because it has the property that it encloses every interior point of a square, and yet it is a closed curve whose area is less than half that of the square. This program draws successive generations to illustrate how the Sierpinski curve, which is the limit of these polygonal drawings, is constructed. 1 REM Sierpinski Curve 10 INPUT"MODE"0 15 INPUT"SIZE"K 20 CLEARO 30 S=5 40 J=l 50 FOR I=1 TO 5 60 J=J*2;D=K/J/4 70 X=K-5*D; Y=K-2*D 80 T=l; MOVE X,Y 90 X=X+D; A=J; B=J; GOTO s 100aIF A=J AND B=J GOTO z 110sP=J; Q=a; R=B 120vIF P<2 goto z 130 if p="2 GOSUB o; GOTO a 140" p="P/2 150 IF Q<p OR P+1<Q GOTO n 170 IF R<P OR P+1<R GOTO n 190 GOSUB c; GOTO a 200nIF Q">=p THEN Q=Q-P 210 IF R>=P THEN R=R-P 220 GOTO v 230zREM end of loop 240 FOR N=1 TQ 1000;NEXT 250 CLEARO 260 NEXT I 270 END 1000cGOTO(1000+100*T) 1100 X=X+D 1105 PLOTS,X,Y 1110 X=X+D;Y=Y+D;PLOTS,X,Y 1120 Y=Y+D;B=B+1;T=4;RETURN 1200 Y=Y-D 1205 PLOTS,X,Y 1210 X=X+D;Y=Y-D;PLOTS,X,Y _81_
1220 X=X+D;A=A+1;T=1;RETURN 1300 X=X-D 1305 PLOTS,X,Y 1310 X=X-D;Y=Y-D;PLOTS,X,Y 1320 Y=Y-D;B=B-1;T=2;RETURN 1400 Y=Y+D 1405 PLOTS,X,Y 1410 X=X-D;Y=Y+D;PLOTS,X,Y 1420 X=X-D;A=A-1;T=3;RETURN 2000oGOTO(2000+100*T) 2100 X=X+D;PLOTS,X,Y 2110 X=X+D;Y=Y+D;PLOTS,X,Y 2120 X=X+D;Y=Y-D;GOTO 1305 2200' Y=Y-D;PLOTS,X,Y 2210 X=X+D;Y=Y-D;PLOTS,X,Y 2220 X=X-D;Y=Y-D;GOTO 1405 2300 X=X-D;PLOTS,X,Y 2310 X=X-D;Y=Y-9;PLOTS,X,Y 2320 X=X-D;Y=Y+D;GOTO 1105 2400 Y=Y+D;PLOTS,X,Y 2410 X=X-D;Y=Y+D;PLOTS,X,Y 2420 X=X+D;Y=Y+D;GOTO 1205 Description of Program: 50 Plot five generations l000-1420 Plot centre square 2000-2420 Not a centre square Variables: A,B - Coordinates of current square D - Number of cells in a quarter of a square J - Number of squares in picture K - Resolution of screen 0 - Graphics mode S - Argument for PLOT statement T - Angle in units of 90 degrees. X,Y - Current drawing position Program size: 1047 bytes Sample plot: sierpinski.png _82_


11.5.2 Three-Dimensional Plotting The following program will plot a perspective view of a three-dimensional object or curve as viewed from any specified point in space. The program is simply provided with a subroutine giving the coordinates of the object to be drawn, or the equation of the curve. The program below plots a perspective view of the curve 1/(1+x^2+y^2) for a range of values of x and y. The function has been scaled up by a factor of 300 to bring the interesting part of the curve into the correct range. The program is provided with an equation of the curve, specifying z (the vertical axis) in terms of x ,ind y (the two horizontal axes), and the view position. It Erojccts every point on the surface onto a plane perpendicular to the line joining the view position to the origin. The example given here draws line of equal y, and the surface is drawn as if viewed from the point x=30, y=40, z=8; i.e. slightly above the surface. 1 REM Three-Dimensional Plotting 50 L=30;M=40;N=8 110 Z=O;CLEAR4 120 A=08000;B=09800 130 FORJ=A TO B STEP4;!J=-1;N. 150 S=L*L+M*M;GOS.s;R=Q 160 S=S+N*N;GOS.s;S=L*L+M*M 170 T=L*L+M*M+N*N 200 F.U=-20TO20 210 V=-20;GOS.c;GOS.b 220 F.V=-19TO20;GOS.c;GOS.a;N.;N. 230 END 400sQ=S/2 410 DOQ=(Q+S/Q)/2 415 U.(Q-1)*(Q-1)<S AND(Q+1)*(Q+1)>S 420 R. 500 REM DRAWTO(UfVfW) 510aZ=3 520k>0=T-U*L-V*M-W*N 530 C=T*(V*L-U*M)*4/(R*0)+128 540 D=96+3*Q*(W*S-N*(U*L+V*M))/(R*0) 560 PLOT(2+4),C,D;Z=O;R. 600cW=300/(10+U*U+V*V)-10;R. Description of Program: 50 Set up view position 110 Set move mode, and clear screen 120-130 Invert screen 150-170 Calculate constants for linear projection 200-230 Scan X,Y plane evaluating function and plotting projected lines. 400-420 s: Square root routine (see Section 5.2.2). 500-560 a: Calculate projected position of next point and move to it (Z=O) or draw to it (Z=3) 600 c: Function for evaluation Variables: A - Display area start B - Display area end C,D - Coordinates of projected point J - Display location to be cleared Q,r,s,t - Constants for projection U,V - Scan variables _83_
W - Function value Program size: 491 bytes Sample Plot 3dplot.png


11.6 Plotting Hex Characters In the higher graphics modes, modes 1 to 4, characters cannot be plotted on the screen directly but it is fairly simple to draw characters using the graphics statements. The following simple routines will draw the hex characters 0 to F, with any desired scaling, and with an optional slope. The routines are useful for labelling graphs drawn in the higher-resolution qraphics modes. Routine p plots a single hex character; routine q plots two hex characters. The routine is demonstrated by drawing random hex characters in a circle. 1 REM Plotting Hex Characters 10 N=TOP; !N=06E3E4477; N!4=0467B6B4D 12 N!8=f795F4F7F; N!12=01B3B7C33 20 V=2; H=2; S=O 25 CLEAR 0 30 X=30; Y=O 40 MOVE (32+X),(24+Y) 50 X=X+Y/6;Y=Y-X/6 60 A=ABSRND&OF 70 GOSUBp 90 GOTO 40 1000qREN Plot B as 2 hex digits 1010 A=B/16; GOSUB p 1020 A=B&#F 2000pREM Plot A in hex 2001 REM uses:A,H,J,K,L,N,Q,V 2010 Q=n?a 2020 FOR J=1 TO 7 2030 K=(2-J%6)%2;L=(2-(J'-1)%4)%2 2040 PLOT(Q&1),(L*H+K*S),(K*V) 2050 Q=Q/2; NEXT J 2060 PLOTO,((H+2)/2),0; RETURN _84_
Description of Program: 10-12 Set up plotting statements for the 16 characters. 20 Scales for letters 30-50 Move X,Y around a circle 60-70 Plot random character 1000-1020 q: Plot low-order byte of B as two hex digits 2000-2060 p: Plot low-order hex digit of A in hex Variables: A - Hex digit to be plotted B - Byte to be plotted H - Horizontal scaling N - Vector containing character plotting statements Q - Next plot statement; low-order bit determines whether to draw or move. S - Slope factor V - Vertical scaling X,Y - Coordinates of point on circle. Program size: 457 bytes Vector: 16 bytes


11.7 Animated Graphics The graphics statements are optimised for speed. For example, to draw a diagonal across the screen using: MOVE O,O ; DRAW 255,191 takes under 40 msec. The following program uses animated graphics to display a clock whose hands move to show the correct time. The hands are drawn using the statement PLOT 6, and the same statement is repeated to remove each hand’s old position before drawing its new position. The clock keeps accurate time by executing the WAIT statement: 1 REM Clock 10 CLEAR4;E=128;F=96 15 J=71;K=678;Q=100;R=OB001 20 X=O;Y=8000;G=90 30 MOVE(X/Q+E),(Y/Q+F) 40 FORL=O TO 59 45 IF L%5<>0 GOTO c 50 DRAW(X/G+E),(Y/G+F) 55 MOVE(X/Q+E),(Y/Q+F) 60cGOSUBi;GOSUBp 68 NEXTL 70 X=O;Y=5000;S=O 72 DO A=O;B=6600 80 FOR H=O TO 4 82 GOSUBh;C=X;D=Y;X=A;Y=B 84 FOR M=O TO 11 85 GOSUBh;A=X;B=Y 87 X=O;Y=7000 88 IF ?R<>OFF GOTO b 90 FOR L=O TO 59 110 GOSUB s 120 FOR N=S TO 55;WAIT;NEXT N 130 S=O 140 GOSUBs;GOSUBi 150 NEXT L 155bX=A;Y=B 160 GOSUBh;GOSUBi _85_
170 NEXT M 175 A=X;B=Y;X=C;Y=D 180 GOSUBh;GOSUBi 200 NEXT H; UNTIL 0 399 REM 400h510VE E,F 410 V=X/2/Q;U=Y,¿2’Q;W=Vj5;T=U/5 415 WAIT 420 PLQT6,(V-T+E),(U+W+F) 430 PLOT6,(X/Q+E),(Y/Q+F) 440 PLOT6,(V4T+E),(U-W+F) 450 PLOT6,E,F;S=S+5;RETURN 500iWAIT;X=X+J*Y/K 510 Y=Y-J*X/K;S=S+1;RETURN 600sMOVE E,F 620pWAIT;PLOT6,(X/Q+E),(Y/Q+F) 630 S=S+1;RETURN Description of Program: 40-68 Draw clock face 80-84 Do hours and minutes 88 If shift key down miss out seconds 90-150 Do seconds 120 Use up remainder of each second 400-450 h: Draw hour/minute hand from centre of screen to X,Y 500-510 i: Increment X,Y one sixtieth of way around circle. 600 s: Draw second hand 620-630 p: Plot to point X,Y Variables: A,B - Coordinates of tip of minute hand C,D - Coordinates of tip of hour hand E,F - Coordinates of centre of screen H - Twelves of minutes counter J,K - Incremental variables; J/K = 2*PI/60 approx. L - Seconds counter M - Minutes counter N - Counter Q - Scaling factor R - Address of shift key S - Sixtieths of a second used out of current second X,Y - Coordinates on screen scaled by Q Program size: 806 bytes _86_
Sample Plot: clock.png To set the correct time hold the shift key down after typing RUN, and release it when the hour and minute hands are in the correct positions.


11.8 Plotting in BASIC To illustrate how the plotting statements work, the following BASIC programs will plot points on the screen in the different graphics modes without using PLOT, DRAW, or MOVE.


11.8.1 Plotting and Testing Points in Mode 0 The following BASIC program will plot a point in the graphics mode 0; the main program sets up a vector V which contains bytes with a single bit set to denote the bit to be plotted. Subroutine p plots a point at the coordinates X and Y. 1 REM Plot in Mode 0 10 DIM V(5) 20 !V=#04081020; V!4=#102 100 REM Plot point at X,Y 110 REM Changes: P; Uses V,X,Y 120pP=X/2+(47-Y)/3*32+#8000 130 ?P=?P\IV?(X&1+(47-Y)%3*2);RETURN Using this method it is possible to determine the state of any point on the screen, as well as actually plotting points. For example, changing line 130 to: 130 Q=(?p&(v?(x&1+(47-y)%3*2))<>0) uses Q as a logical variable whose value is set to ’true' if the point X,Y is set, and to 'false' if the point is clear. Note that the screen should be cleared by writing 040 in every location (or with the statement CLEAR 0) before plotting in graphics mode zero with this routine. _87_


11.8.2 Plotting in Higher Graphics Modes To set the ATOM to a higher graphics mode the following character should be stored in location #BOOO: Mode: Value: 0 #00 1a #10 1 #30 2a #50 2 #70 3a #90 3 #B0 4a #DO 4 #F0 This operation is performed automatically for modes 0, 1, 2, 3, and 4 by the CLEAR statement. Modes la, 2a, 3a, and 4a are colour graphics modes; see section 11.9 below. To illustrate plottivg in the higher modes the following BASIC program will plot a point on the screen at the coordinates X,Y in the highest-resolution graphics mode: 10 DIM V(7) 20 !V=410204080; V!4=#1020408 30 ?#B000=#F0 100 REM Plot point at X,Y 110 REM Changes: P; Uses: V,X,Y 1OOpP=X/8+(191-Y)*32+#8000 102 ?P=?\V?(X&7);RETURN Again the program can be modified to test the state of points of the screen.


11.9 Colour Graphics The ATOM provides three additional graphics modes which provide graphics in four selectable colours up to a maximum definition of 128x192. These modes are known as la, 2a, 3a, and 4a. The BASIC's PLOT, DRAW, and MOVE statements can be used in the 4-colour modes provided that a point-plotting routine, written in assembler, is provided to replace the black-and-white point plotting routines. Alternatively the COLOUR statement, provided in the extension ROM, can be used; see Section 22.2. The address of the point-plotting routine used by PLOT, MOVE, and DRAW is stored in RAM at 03FE and 03FF. The following information is passed down to the point-plotting routine in zero page: Location: Function: 5A X coordinate - low byte 5B " " high byte 5C Y coordinate - low byte 5D " " high byte 5E 1: set bit, 2: invert bit, else, clear bit. 5F Free for workspace 60 " " The following BASIC program demonstrates how an assembler point-plotting routine can be provided to give four-colour plotting in graphics mode 4a, the highest-resolution colour graphics mode: _88_
10 REM 4-Colour Plot 12 GOSUB 400 16 CLEAR4;?4BOOO=ODO 18 ?03FE=Q;?03FF=Q&OFFFF/256 30 FOR J=O TO 64 STEP 2 40 ?C=J%3*4;MOVE J,O 50 DRAW 127,J;DRAW(127-J),191 60 DRAW 0,(191-J);DRAW J,O 70 NEXT J 80 END 400 DIM V(11),C(0),P(-1),Q(-1) 420 !V=#01041040;V!4=002082080;V!8=#030C30CO 430 P.$21 508[ 510 LDA@O;STA #5F 520 LDA#5C;LSR A;ROR #5F 530 LSRA;ROR#5F;LSRA;ROR#5F 540 STA#60;LDA#5A;LSRA;LSRA 550 CLC;ADC#5F;STA05F 560 LDA#60;ADC@#80;STA#60 570\#5f AND #60 CONTAIN ADDRESS 580 LDA05A;AND@3;CLC;ADCC;TAY 590 LDX@0;LDAV,Y;ORA(#5F,X) 600 STA(#5F,X);RTS 610] 620 P.$6 630 RETURN Description of Program: 12 Assemble point plotting routine 16 Clear display; set mode 3a 18 Change point plotter vector 30-70 Demonstration program; curve stitching in 4 colours 400 Set up variable space 420 Vectors for three colours 430 Disable assembler listing 508-610 Assembler point-plotter program 620 Turn screen back on Variables: C - Colour: 0, 4, or 8. P - Location counter Q - Address of point-plotting routine V - Vectors for setting bits Program size: 558 bytes Vectors: 13 bytes Note that the routine only sets bits, and plots in three colours - the fourth colour being the background colour. It would be a simple matter to modify the routine so that it was able to set or unset bits; i.e. plot in the background colour. _89_
_90_

12 What to do if Baffled

This section is the section to read if all else fails; you have studied your program, and the rest of the manual, and you still cannot see anything wrong, but the program refuses to work. There are two types of programming errors; errors of syntax, and errors of logic.


12.1 Syntax Errors Syntax errors are caused by writing something in the program that is not legal, and that is therefore not understood by the BASIC interpreter. Usually this will give rise to an error, and reading the description of that error code in Chapter 27 should make the mistake obvious. Typical causes of syntax errors are: 1. Mistyping a digit '0' for a letter '0', ard vice-versa. E.g.: FOR N=1 TO 3 2. Mistyping a digit '1' for a letter 'I’, and vice-versa. E.g.: 1F J=2 PR1NT "TWO" 3. Forgetting to enclose an expression in brackets when it is used as a parameter in a statement. E.g.: MOVE X+32,Y In some cases a syntax error is interpreted as legal by BASIC, but with a different meaning from that intended by the programmer, and no error message will be given. E.g.: MOVE 0,0 was intended to move to the origin, but in fact moves to some coordinates dependent on the value of the variable 0.


12.2 Logical Errors Errors of logic arise when a program is perfectly legal, but does not do what the programmer intended, probably because the programmer misinterpreted something in this manual, or because a situation arose that was not forseen by the programmer. Common logical errors are: l. Unitialised variables. Remember that the variables A-Z initially contain unpredictable values, and so all the variables used in a program should appear on the left hand side of an assignment statement, in an INPUT statement, dimensioned by a DIM statement, or as the control variable in a FOR...NEXT loop, at least once in the program. These are the only places where the values of variables are changed. 2. The same variable is used for two purposes. It is very easy to forget that a variable has been used for one purpose at one point in the program, and to use it for another purpose when it was intended to save the variable's original value. It is good practice to keep a list of the variables used in a program, similar to the list given after _91_
the application programs in this manual, to avoid this error. 3. Location counter P not set up when assembling. The value of P should be set before assembling a program to the address of an unused area of memory large enough to receive the machine code, and P should not be used for any other purpose in the program. 4. Graphics statements used without initialising graphics. The CLEAR statement must precede use of any graphics statements. 5. Assigning to a string variable and exceeding the allocated space. Care should be taken that enough space has been allocated to string variables, with DIM, to receive the strings allocated to them. 6. Assigning outside the bounds of an array or vector. Assigning to array or vector elements above the range dimensioned in the DIM statement will overwrite other arrays, vectors, or strings.


12.3 Suspected Hardware Faults This section deals with faults on an ATOM which is substantially working, but which exhibits faults which are thought to be due to hardware faults rather than programming faults. Hardware fault-finding details are provided in the Technical Manual; this section describes only those hardware problems that can be tested by running software diagnostics.


12.3.1 RAM Memory Faults The fallowing BASIC program can be used to verify that the ATOM's memory is working correctly: 1 REM MEMORY TEST 10 INPUT"FROM"A," TO"B 20 DO ?12=0; R=!8 30 FOR N=A TO B STEP4;!N=RND; NEXT N 35 ?12=0; !8=R 40 FOR N=A TO B STEP4 50 IF !N<>RND PRINT'"FAIL AT "&N' 60 NEXT N 70 P." OK"; UNTIL 0 The first address entered should be the lowest address to be tested, and the second address entered should be four less than the highest address to be tested. For example, to test the screen memory enter: >RUN FROM?#8000 TO?#81FC The program stores random numbers in the memory locations, and then re-seeds the random-number generator and checks each location is correct.


12.3.2 ROM Memory Faults The BASIC interpreter, operating system, and assembler, are all contained in a single 8K ROM, and as all ROMs are thoroughly tested before despatch it is very unlikely that a fault could be present. However, if a user suspects a ROM fault the following program should be entered and run; the program obtains a 'signature' for the whole ROM, this signature consistinq of a four-digit hexadecimal number. The program should be run for each 4K half of the ROM. _92_
1 REM CHC Signature 10 INPUI' "LOOM ADDRESS", P 20 C=O;Z=OFFFI';Y=#2D 30 FOR 0=0 TO #FFF 35 A=P?Q 40 FOR B=l TO 8 60 C=C*2+A&1;A=A/2;IFC>Z C=C:Y;C=C&Z 80 NEXT B; NFXT Q 110 PRINT "SIGNATURE IS" &C' 120 END Program size: 213 bytes Sample run: >RUN PROM ADDRESS?#COOO SIGNATURE IS D67D >RUN PROM ADDRESS?#FOOO SIGNATURE IS E386 The program takes about 6 minutes to run, and if these signatures are obtained the ROM is correct. The Atom extension ROM, described in chapter 22, can be tested by giving the reply #DOOO to the prompt. It should give a signature of AAA1.


12.3 Programming Service If all else fails, owners of an ATOM may make us¿ of the free Programming Service provided by Acorn. To ensure a rapid reply to any queries the special Programming Service Forms, supplied with the ATOM, must be used to submit the problem New forms will be supplied with the reply to any queries, or on request. All reports should be accompanied by a full description of the problem or fault, and the occasions when it occurs. Please also enclose a stamped addressed envelope for the reply. A program should be supplied which illustrates the problem or suspected fault. This program should preferably be only four or five lin s long, and should be written in the space provided on the Programming Service Form, with any spaces in the original carefully included. If the problem or fault is only exhibited by a longer program the report form should be accompanied by a cassette tape recording of the program, and the title of the file on the tape should be entered on the form. The cassette will be returned with the reply. _93_