With trepidation in this apparently exalted company I would point out that there is a critical difference to the softwares' processing of data coming from a source which has self tested or otherwise signalled to the software 'FAILURE', compared to the processing of a parameter which is erroneous.
In the absence of a comparitive function and/or detection of values outwith a 'normal' range the eromeous RA values would be processed in the same way as 'true' data. e.g. negative altitude value,as the software would have no way of detecting that the data was incorrect.
The very valuable tests you are conducting must take cognisance of this,and contributors should be very careful only to use 'failure' when the instrument has outputted a failure signal or the software has otherwise detected invalid data,flagged the source as 'failure' and processed/ignored the data accordingly.
The softwares' reaction to 'failure' c/f 'erroneous'(undetected) will be poles apart.
Apologies for being so pedantic,but unfortunately there is nothing so pedantic as control software,miss out e.g. a closing bracket and it bites your @rse). Where DID I put that opening bracket?
