Full Idea
For the simple regularity theorist, the function ought to be a gappy one, leaving out values not actually instantiated; …one function would fit the actual points on the graph as well as any other.
Gist of Idea
If laws are just instances, the law should either have gaps, or join the instances arbitrarily
Source
Alexander Bird (Philosophy of Science [1998], Ch.1)
Book Reference
Bird,Alexander: 'Philosophy of Science' [UCL Press 2000], p.32
A Reaction
The 'simple' theorist says there is nothing more to a law than its instances. Clearly Bird is right; if the points line up, we join them with a straight line, making counterfactual assumptions about points which were not actually observed.