I am involved in the development of a number of radar systems. At a recent Critical Design Review (CDR) for one of the radar systems there was disagreement regarding whether the signal processing design needed to be finished for the CDR to be passed. Now of course no radar design can be definitivley finalised until it has been manufactured, tested and accepted by a customer: that was not in dispute. It is also understandable that certain algorithm parameters may need tweaking between CDR and Factory Acceptance Test (FAT). I can also accept that the coding of the software could be part of the post-design ie 'manufacture' phase that occurs after CDR, but I was expecting the algorithms that form the signal processing to have been selected (such as whether to use a clutter map or not and whether different types of targets should be processed using different parallel processing chains). I was also expecting to see results showing that the signal processing algorithms had been tested on suitably representative data, perhaps simulated data, to provide confidence that they would meet the required performance status. Am I expecting too much?