During my attempts to test the MK-90 I have written a simple (thus ugly and ineffective) BASIC code for the Gauss Elimination Method. The data (a left-side matrix and a right-side vector) are located in the DATA statement.

REM Data input:

10 DATA 3 1 2 3 2 1 3 3 3 7 5 5 5

20 READ M

25 LET N = M-1

30 DIM A(9,9)

35 DIM B(9)

36 DIM X(9)

40 FOR I = 0 TO N

50 FOR J = 0 TO N

60 READ A(I,J)

70 NEXT J

80 NEXT I

90 FOR I = 0 TO N

95 READ B(I)

100 NEXT I

```
```REM Forward run:

300 FOR J = 0 TO N

310 FOR I = J+1 TO N

320 LET C = A(I,J)/A(J,J)

330 FOR K = 0 TO N

340 LET A(I,K) = A(I,K) - C*A(J,K)

350 NEXT K

360 LET B(I) = B(I) - C*B(J)

370 NEXT I

380 NEXT J

390 LET X(N) = B(N) / A(N,N)

REM Backward run:

400 FOR I = N-1 TO 0 STEP -1

410 LET S = 0

420 FOR J = I+1 TO N

430 LET S = S + A(I,J) * X(J)

440 NEXT J

450 LET X(I) = (B(I)-S)/A(I,I)

460 NEXT I

`REM Results:`

500 PRINT "Results:"

510 FOR I = 0 TO N

520 PRINT X(I)

530 NEXT I

It should work in the old PDP-11 BASIC, and with the old Soviet BASIC (DVK computers, MK-90 and so). It is not suitable for the MK-85 because it supports only 1D fields (and also because it wastes the memory with unnecessary stuff which is a problem for the MK-85).

And, of course, it is compatible with the Bywater BASIC which I recently have compiled for all my main computers (the SGI O2, the NanoNote and for the Linux-based Zauri)