Originally Posted by TwoOneFour
It was essentially a complete scan of the areas which hadn't already been searched.
Not exactly. It was a weighted search of the suspected area. It would eventually have been "complete" if unsuccessful at any point. But it focused on particular areas up-front. Metron, the US search consultants, used a well-known search formula (Bayes' rule) to figure high-probability areas for the search. Here is a simple 3rd-party
precis of the forumal and methodology:
Bayes' rule found AF 447 - Sharon Bertsch McGrayne
Keep in mind, this could in no way be created without prior knowledge of search failures.
Originally Posted by TwoOneFour
Eventually someone decided to stop pussyfooting around with probabilities and, on the assumption it had to be there somewhere, just scour the 10,000sqkm region from one end to the other until they found it.
No, it was not a systematic drag or grid search of the area. Here's a detailed report about how the searches were performed, both early (failed) and later (successful). If you skip to page 17 (fig. 14), you'll see the final search pattern priorities. Compare that to fig. 15 on the following page - the previous search probabilities of a systematic search
a priori. It's not "systematic." It uses the Bayesian method to assign probabilities, the highest of which were to the previously unsearched areas.
http://www.bea.aero/fr/enquetes/vol....h.analysis.pdf