Test Summary

Throughout the writing of the ASTRA program every effort has been made to ensure it performs as intended. All of the subroutines have been individually tested to some degree and in the case of the geometry routines extensive manual checking was performed to ensure as much as possible that its operation is correct. The completed program has then been tested against the results from other programs to check the operation of the program as a whole. It appears to have withstood all these tests and it might therefore be hoped that it can be used with confidence for new predictive work. One note of caution is that much of the testing has been against earlier versions of the STEN program. The testing of the differential rotation against the published results of Richer and Padman does not really negate this as their results check only that the general line shape for an optically thin cloud works correctly. It does not test the ability of the program to correctly calculate the output from optically thick rotating clouds. A important test for the future will be to compare the output for more complex models against a completely independent program such as a Monte Carlo simulation. Nonetheless the successful testing of ASTRA against STEN is extremely encouraging as STEN itself has been successfully tested by Jeremy Yates (private communication) against a completely independent program that uses the finite difference equation method (see section 3.4).

The testing here does however give enough confidence to proceed to try and apply the ASTRA program to some new models that are a more realistic simulation of the conditions that are believed to exist in star formation regions.