Inside Atari Microsoft BASIC: A First Look
Jim Butterfield
Atari's long-awaited Microsoft BASIC is here at last. Jim Butterfield, an expert on the 8K Microsoft BASIC used on other machines, begins the documentation of the complex inner workings of Atari Microsoft BASIC.
It's a big BASIC. It occupies 18K of RAM, which means there's a lot of code in the interpreter. It also does some new things. Single versus double precision arithmetic, for example, calls for a dramatic rearrangement of the floating accumulators and of the way variables are stored as compared to the better-known 8K Microsoft BASICs.
With the expanded features come new techniques to be mastered. I wince when PRINT 10/4 yields an answer of 2 (to get 2.5, you must force floating point with PRINT 10/4.0).
The Architecture
The following discussion assumes that users have had some exposure to the mechanics of other Microsoft BASICs.
Your BASIC program will be stored right behind the interpreter (Hex 6980 and up). It's the usual format: two-byte forward chain to the next BASIC line, two-byte BASIC line number, the line itself (tokenized) and finally a zero byte to flag end-of-line. The end-of-program is identified by zero-bytes in the forward chain location.
Variables come behind your program – check the address in hex 84 and 85, or PRINT PEEK(132) + PEEK(133) * 256 – but storage is fairly complex. The first two bytes are the first two characters of your variable name, but with many bits stripped away and replaced with "variable type" bits: don't be surprised if your variable A ends up with the name stored as value 1 rather than the ASCII 65 which corresponds to A. The third byte is the length of this variable entry. Now we have a messy bit: if you have a long variable name such as PLUGH the extra letters (UGH) are stored starting at the fourth byte. Finally, the value itself.
The following memory map is a brief list of the locations I have spotted while looking around. It's far from complete; but those who would like to rummage around will find it handy.
| 0080-0081 | 128-129 | Pointer : Start-of-Basic |
| 0082-0083 | 130-131 | Pointer |
| 0084-0085 | 132-133 | Pointer : Start-of-Variables |
| 0086-0087 | 134-135 | Pointer : Start-of-Arrays |
| 008A-008B | 138-139 | Pointer : String storage (moving down) |
| 0094-0095 | 148-149 | Current data pointer |
| 00AE-00C5 | 174-197 | CHRGET subroutine |
| 00B4 | 180 | CHRGOT entry point |
| 00B5-00B6 | 181-182 | Basic pointer within subroutine |
| 00C7-00C8 | 199-200 | Variable pointer for FOR/NEXT |
| 00CB | 202 | $98 = READ, $40 = GET, 0 = INPUT |
| 00CD | 204 | Default DIM flag |
| 00D1 | 209 | Accum sign compare, #1 vs #2 |
| 00D2 | 210 | Accum #1 Low order (rounding) |
| 00D4 | 212 | Variable name length |
| 00D5-00D8 | 213-216 | Utility Pointer area |
| 00DC-00E4 | 220-228 | Misc numeric work area |
| 00E5 | 229 | Accum #1 precision flag |
| 00E6 | 230 | Accum #1 : Sign |
| 00E7 | 231 | Accum #1 : Exponent |
| 00E8-00EE | 232-238 | Accum #1 : Mantissa |
| 00F0 | 240 | Accum #2 : Sign |
| 00F1 | 241 | Accum #2 : Exponent |
| 00F2-00F8 | 242-248 | Accum #2 : Mantissa |
| 00F9-00FF | 249-255 | Product area for multiplication |
| 0480- | 1152- | Variable name setup area |
| 1F00-697F | 7936-27007 | Microsoft Basic Interpreter |
| 6980- | 27008- | Basic program staging area |
Return to Table of Contents | Previous Section | Next Section
