Assuring safety-related data - how do we decide the data is correct?

'Data' can cover a multitude of forms - descriptions of physical measurements, configuration details, through to material which looks very like software. Data is particularly malleable, and often gets converted from one format to another, potentially including transmission through untrusted media. This talk describes some approaches for categorising data and the different problems that each category poses for assurance together with some techniques for checking data. This can also include cases where the solution is not to trust the data!