In working through functional safety examples with @tobiaskaestner there's a desire to have an integritymethod applied to a relationship to enable appropriate audit methodology.
Open question is what should this be computed from: from & tos?
Should we have something like a package verification code (hashes of hashes - starting with from, concatenated with all tos).
What should do about properties in the relationship itself? Ideally they should be included, but we can't include integrity method itself. ;-)
The evidence only holds if the test specification to the test results belongs relates to the software requirement at the correct version.
If you change the requirement, the integity should be able to be detected as invalidate; and a new relationship minted with the appropriate integrety checks.
@nicpappler - do you agree?
In working through functional safety examples with @tobiaskaestner there's a desire to have an integritymethod applied to a relationship to enable appropriate audit methodology.
Open question is what should this be computed from: from & tos?
Should we have something like a package verification code (hashes of hashes - starting with from, concatenated with all tos).
What should do about properties in the relationship itself? Ideally they should be included, but we can't include integrity method itself. ;-)
The evidence only holds if the test specification to the test results belongs relates to the software requirement at the correct version.
If you change the requirement, the integity should be able to be detected as invalidate; and a new relationship minted with the appropriate integrety checks.
@nicpappler - do you agree?