Currently, E2E works by comparing running the simulation and comparing the results against expected values. If it detects a difference, it flags it. This approach does not account for the cases where we expect a certain set a variables to be different than the expected variables. This is particularly important when we have generated the E2E filter using version X and we want to compare those against the results of version X+. If an SME knows for sure that some variables are expected to be different in these two versions, we need to let them express that. Two thoughts on how do this: (1) introduce a flag next to each variable so it can be flagged as must be different, or (2) or make a file that hosts the variables that are expected to be different. Or come up with a different way. It is important to keep this accessible and as user friendly as possible.
Note that while an SME can predict a variable must change, they don't always know the new value. For this reason, this is a binary signal rather a scalar value. So for example, the SME tells RuFaS that milk yield must be different; but doesn't tell what's the new value.
As current behavior, we need to compile a list a variables that have a different value than what we have in E2E filters, so an SME can evaluate them.
Currently, E2E works by comparing running the simulation and comparing the results against expected values. If it detects a difference, it flags it. This approach does not account for the cases where we expect a certain set a variables to be different than the expected variables. This is particularly important when we have generated the E2E filter using version
Xand we want to compare those against the results of versionX+. If an SME knows for sure that some variables are expected to be different in these two versions, we need to let them express that. Two thoughts on how do this: (1) introduce a flag next to each variable so it can be flagged asmust be different, or (2) or make a file that hosts the variables that are expected to be different. Or come up with a different way. It is important to keep this accessible and as user friendly as possible.Note that while an SME can predict a variable must change, they don't always know the new value. For this reason, this is a binary signal rather a scalar value. So for example, the SME tells RuFaS that milk yield must be different; but doesn't tell what's the new value.
As current behavior, we need to compile a list a variables that have a different value than what we have in E2E filters, so an SME can evaluate them.