I have identified what may be a small bug that does not create errors, but may result in humans making errors when using openLCA. Here is how to reproduce it:
1) For a given flow, enter a logarithmic uncertainty distribution with a Geometric mean = amount. Process becomes starred (indicates there was a change). Save. Everything's ok.
2) For a given flow, enter a logarithmic uncertainty distribution with a Geometric mean <> amount. Process becomes starred (indicates there was a change). Save. Note that there is now a discrepancy between the amount and the geometric mean (the latter used in Monte Carlo, the former in a deterministic calculation), but that is not openLCA's fault and this may be what the user wants. Everything's ok.
3) Open the Data Quality Entry, change the pedigree scores, and choose "Use as uncertainty value". This will update the Geometric Mean to the Amount value. This may surprise some users, a pop-up may be useful, but this is not what I wanted to talk about. Process becomes starred (indicates there was a change). Save. Everything's ok, sort of.
4) Change the "Geometric mean" value again and save. Then, open the Data Quality Entry window. Do not change anything. Press Use uncertainty value. This changes the Geometric mean (sets it to the Amount value). However, the process is not starred (user does not know something has changed). This is what I wanted to bring to your attention.
Thanks for looking into this.