Chapter 2: The Structure of the Interpreter

C000	BASIC	--- BASIC

F000	BASIC COS

D000	BASIC	--- FLOATING POINT ROM

A000	BASIC	--- UTILITY ROM

Programs are stored in memory as a series of strings, which in the Atom are normally begun at #2900. Address 2900 contains and 0D which means start of program. Each line in the program consists of a two-byte line number (stored as hex), followed by the actual ASCII code for whatever you typed in. At the end of each line is an 0D, and the end of the program is marked by an FF (thus a program always ends in 0D FF). A program consisting only of 20 PRINT "HELLO";END would look like this if we did an ASCII dump starting at 2900:

0D 0C 14 P R I N T " H E L L O " E N D 0D FF

P.&TOP would give 2915, since this is the next memory location after the FF at the end of the program.

Strings being interpreted, either in direct mode or as a program being run, are first checked by the C000 BASIC interpreter. If they are valid, a match with the word in the string is found in the ROM, and the appropriate routines are called for the execution of the word.

If the C000 interpreter can't find a match for the string, then it passes control over to the F000 BASIC interpreter. Again, valid matches are sought, and executed if one is found.

If the F000 interpreter can't resolve the string, then normally this would mean that an erroneous string is present, and an error routine is called. However, before giving up all hope, a simple test is made which looks for the signature of a ROM at D000 (the floating-point ROM), and if the ROM is present, then the string is passed over to it for interpretation.

By this means the Atom can work with or without the floating-point ROM installed, and when one is plugged in, the machine is able to detect that it is there.