80-Bus News

entered using a text editor such as ED (if you must), WordStar or PEN. From the source code program (which must have the extension type .BAS), it produces an .INT file in which, for example, ‘PRINT’ is stored as 1Ah. The logic behind this is that it takes a shorter time to interpret 1 byte such as 1A when the semi-compiled program is run than it does to deal with 5 letters as in PRINT. To use the language, you need to:

1. Write a CBASIC source program.
2. Use CBASIC to produce the intermediate (.INT) file.
3. Use CRUN to execute the .INT file.

CRUN is a run-time interpreter which takes the INT file produced by CBASIC after checking the source code for errors, and converts it to machine code. The pseudo-code compiler/​run-time interpreter combination is a compromise between the readily alterable but slow-running interpreted and the fast-running but not easily alterable compiled versions of BASIC.

The same procedure could be used for EBASIC, which, I believe, is available as Public Domain software at a very modest charge from the CP/M Users Group.

CBASIC appears to be more cumbersome in use than MBASIC, but it has a number of very useful features which the latter lacks. The major use for it has been in the field of business applications software where its speed of execution and facilities have been most favoured. Unfortunately, it is only available for disk-based CP/M systems.

The language takes a little getting used to and I would not recommend it to anyone who has less than a good grasp of BASIC: until one is reasonably familiar with its syntax and little peculiarities, it is best not to use some of the more advanced features. In the following pages, I have attempted to provide some information on CBASIC and MBASIC (disk version) statements and commands, where they differ or where they are peculiar to CBASIC. CBASIC has an excellent, if expensive User Guide by A. Osborne, G. Eubanks and M. McNiff, published by Osborne/­McGraw-Hill in 1981. It is rather costly at £12 (in 1983) for 215 pages but is essential for the serious user. The Guide deals with version 2.07 of the compiler; I have used version 2.02 only, but I have not found any major problems or incompatibilities. The version number of the run-time interpreter CRUN used in the book is not stated but mine was 2.04.

Data types

Both BASICs support integer and floating point (real) numbers, with exponential notation for the latter. The range for integers is the same in both (+32767 to −32768), but real numbers are stored by CBASIC with an accuracy of 14 digits, compared with 7 for single precision and 17 for double precision MBASIC. Exponential range for MBASIC is 2.9387E-38 to 1.70141E38, whereas CBASIC has a range from 9.9999999999999E62 to 1.0E-64, which could be very useful!! Both BASICs allow hexadecimal notation, but CBASIC allows binary and MBASIC permits octal notation in addition. All three are, of course, special ways of representing integer numeric data in BASIC.

Names of variables may be up to 31 characters long in CBASIC; although in MBASIC they can be of any length, only the first two alphanumeric characters are recognized; some characters are not allowed but who wants to use ★,!,@ or & in a variable name? Both BASICs use the $ sign to denote string variables, and integers may be explicitly declared. This is done in CBASIC by adding a % sign to the end of the variable name (with not more than 30 preceding characters}; in MBASIC, the DEFINT statement does this. The ability to use variable names like ‘microcomputer’, ‘collywobble’ or ‘knightsbridge’ may appeal to the whimsical but unfortunately, ‘super­cali­fragi­listi­expi­ali­docious’ is too long and would be truncated! Variable names should be distinct and descriptive, which will help you understand the flow of each program you write.

DIMension statements have virtually the same functions in both BASICs; a little appreciated fact is that it is possible to use a constant or a variable to declare the highest subscript for an array or vector:

Z%=50
DIM SORTCODE$(Z%),​HOUSENUMBER(Z%)

or

Z%(1)=50 Z%(2)=5 DIM SORTCODENUMBERS$(Z%(1),​Z%(2))

Statement construction

Line numbers are optional in CBASIC except where the line is referenced by a statement elsewhere in the program. An additional and somewhat bizarre feature is that the numbers can be in any format — integer, real or exponent, but there must be less than 31 digits.

A single CBASIC statement can be spread over several lines by the use of the backslash (\) and lower case letters are converted to upper case unless the ‘D’ toggle is used.

IF REPT = 0 \
THEN GOTO 999

There seems to be no limit to the length of a statement but keep it within reasonable bounds otherwise your source will look like one of those