Appendix C: A description of the ZX Spectrum / The BASIC

Appendix C

Part 1: A description of the ZX Spectrum for reference

In the original paper version of this manual, the first section of this appendix was a repeat of that part of Chapter 1 concerning the keyboard and display.

The display window

Each character position has attributes specifying its paper (background) and ink (foreground) colours, a two-level brightness, and whether it flashes or not. The available colours are black, blue, red, magenta, green, yellow and white.

The edge of the screen can be set to any of the colours using the border statement.

A character position is divided into 8×8 pixels and high resolution graphics are obtained by setting the pixels individually to show either the ink or paper colour for that character position.

The attributes at a character position are adjusted whenever a character is written there or a pixel is plotted. The exact manner of the adjustment is determined by the pnnting parameters. of which there are two sets (called permanent and temporary) of six: the PAPER, INK, FLASH, BRIGHT, INVERSE and OVER parameters. Permanent parameters for the top part are set up by PAPER, INK, etc, statements, and last until further notice. (Initially they are black ink on white paper with normal brightness, no flashing, normal video and no overprinting). Permanent parameters for the bottom part use the border colour as the paper colour, with a black or white contrasting ink colour, normal brightness, no flashing, normal video and no overprinting.

Temporary parameters are set up by PAPER, INK, etc, items, which are embedded in PRINT, LPRINT, INPUT, PLOT, DRAW and CIRCLE statements, and also by PAPER, INK, etc control characters when they are printed to the television - they are followed by a further byte to specify the parameter value. Temporary parameters last only to the end of the PRINT (or whatever) statement, or, in INPUT statements, until some INPUT data is needed from the keyboard, when they are replaced by the permanent parameters.

PAPER and INK parameters are in the range 0 to 9. Parameters 0 to 7 are the colours used when a character is printed:

Parameter 8 ('transparent') specifies that the colour on the screen is to be left unchanged when a character is printed.

Parameter 9 ('contrast') specifies that the colour in question (paper or ink) is to be made either white or black to show up against the other colour.

FLASH and BRIGHT parameters are 0, 1 or 8: 1 means that flashing or brightness is turned on, 0 that it is turned off, and 8 ('transparent') that it is left unchanged at any character position.

OVER and INVERSE parameters are 0 or 1:

OVER 0 new characters obliterate old ones
OVER 1 the bit patterns of the old and new characters are combined using an 'exclusive or' operation (overprinting)
INVERSE 0 new characters are printed as ink colour on paper colour (normal video)
INVERSE 1 new characters are printed as paper colour on ink colour (inverse video)

When a TAB control character is received by the television, two more bytes are expected to specify a tab stop n (less significant byte first). This is reduced modulo 32 to n₀ (say), and then sufficient spaces are printed to move the printing position into column n₀. When a comma control character is received, then sufficient spaces (at least one) are printed to move the printing position into column 0 or column 16. When an Enter control character is received, the printing position is moved on to the next line.

The printer

Output to the ZX printer is via a buffer one line (32 characters) long, and a line is sent to the printer

1. when printing spills over from one line to the next,
2. when an Enter character is received,
3. at the end of the program, if there is anything left unprinted,
4. when a TAB control or comma control moves the printing position on to a new line.

TAB controls and comma controls output spaces in the same way as on the television.

The AT control changes the printing position using the column number, and ignores the line number.

The printer is affected by INVERSE and OVER controls (and also statements) in the same way as the screen is, but not by PAPER, INK, FLASH or BRIGHT.

The printer will stop with error B if Esc is pressed. If the printer is absent the output will simply be lost.

Part 2: The BASIC

Numbers are stored to an accuracy of 9 or 10 digits. The largest number you can get is about 10³⁸, and the smallest (positive) number is about 4×10^-39.

A number is stored in floating point binary with one exponent byte e (1≤e≤255), and four mantissa bytes m (1/2≤m<1). This represents the number m×2e^-128.

Since 1/2≤m<1, the most significant bit of the mantissa m is always 1. Therefore in actual fact we can replace it with a bit to show the sign - 0 for positive numbers, 1 for negative.

Small integers have a special representation in which the first byte is 0, the second is a sign byte (0 or FFh) and the third and fourth are the integer in twos complement form, the less significant byte first.

Numeric variables have names of arbitrary length, starting with a letter and continuing with letters and digits. Spaces and colour controls are ignored and all letters are converted to lower-case letters.

Control variables of FOR-NEXT loops have names a single letter long.

Numeric arrays have names a single letter long, which may be the same as the name of a simple variable. They may have arbitrarily many dimensions of arbitrary size. Subscripts start at 1.

Strings are completely flexible in length. The name of a string consists of a single letter followed by $.

String arrays can have arbitrarily many dimensions of arbitrary size. The name is a single letter followed by $ and may not be the same as the name of a string. All the strings in a given array have the same fixed length, which is specified as an extra, final dimension in the DIM statement. Subscripts start at 1.

Slicing: Substrings of strings may be specified using slicers. A slicer can be

1. empty or
2. numerical expression or
3. optional numerical expression TO optional numerical expression

and is used in expressing a substring either by

(a) string expression (slicer)
(b) string array variable (subscript,..., subscript, slicer)
which means the same as
string array variable (subscript . . . , subscript) (slicer)

In (a), suppose the string expression has the value s$.

If the slicer is empty, the result is s$ considered as a substring of itself.

If the slicer is a numerical expression with value m, then the result is the mth character of s$ (a substring of length 1).

If the slicer has the form (iii), then suppose the first numerical expression has the value m (the default value is 1), and the second, n (the default value is the length of s$).

If 1≤m≤n≤the length of s$ then the result is the substring of s$ starting with the mth character and ending with the nth. If 0≤n<m then the result is the empty string. Otherwise, error 3 results.

Slicing is performed before functions or operations are evaluated, unless brackets dictate otherwise.

Substrings can be assigned to (see LET).

If a string quote is to be written in a string literal, then it must be doubled.

Functions

The argument of a function does not need brackets if it is a constant or a (possibly subscripted or sliced) variable.

The following are binary operations:

Functions and operations have the following priorities:

Statements

Note that arbitrary expressions are allowed everywhere (except for the line number at the beginning of a statement).

All statements can be used either as commands or in programs (although they can be more sensible in one than the other). A command or program line can have several statements, separated by colons (:). There is no restriction on whereabouts in a line any particular statement can occur - although see IF and REM.

Appendix B