NOT.Preface:4.3 - Interpret an edit before propagating it

Now a participant asks that the output at q=1 become 4 while the coefficient 2 remains fixed. NOT.6 treats this as a constraint on the rule, not as an isolated replacement of a displayed cell. It gives 2*1 + b = 4, hence b=2. The graph receives the new parameter; the table becomes (0,2), (1,4), (2,6). Its former values were derived outputs, so they are recalculated.

If another participant also requires the old output at q=0 to remain 1, that requirement gives b=1. It conflicts with b=2 under the retained rule. The useful result is the conflicting pair of requirements and the assumption on which it depends. Silently choosing one edit would conceal the decision that the group must make.

An independent measurement record is different again. If its row says that an observed output was 3, a parameter change does not change that past observation. Keep the record and updated prediction separately so that the difference can inform the subject inquiry. Correspondence between representations does not license overwriting every similarly named value.