Beautiful Data

Conclusion

The instrument software did much more than just take pictures and process image data. It also managed motion control with three degrees of freedom for the SSI, and the focus and viewport cover motors in the RAC. The RAC also supported multiple banks of red, blue, and green LEDs to illuminate whatever might be in the robotic arm scoop and create color images. Both the SSI and the RAC incorporated active thermal control, achieved either through the use of special heaters or by intentionally stalling a stepper motor to achieve self-heating. On top of all this, there was the error-checking and fault-recovery code. All in all, it was very busy software.

If I had it to do all over again, I suppose the main thing I would want to see changed would be that the cameras use their own embedded processors rather than rely on the spacecraft CPU. This would have made things much easier all around for everyone. Apart from that, I always felt that there was too much crammed into each of the instrument tasks. In other words, the thermal control should have been a separate task for each camera. This would have greatly reduced the complexity of each of the tasks, albeit at the expense of increasing the overall complexity of the intertask communications. At the outset, however, there wasn't enough evidence to build a compelling case for this, so the design was already firm (not really frozen, just very inflexible) by the time some new thermal requirements popped up that needed to be accommodated. ...