I'm looking for advice on representing bad data.
I am aware of the function ISNAN(), which returns .true. if the argument is an IEEE Not-a-Number. Is NaN a 'recommended' or sensible way of representing bad data? And what's the simplest way of setting a value to NaN without producing a run-time error?
Are there any potential issues I need to watch out for (e.g. reading and writing formatted data containing NaN)?