We can define an inferred node as "successful" if it subsequently becomes an attachment point. It would be interesting to look at how many nodes we infer are successful, and to examine the properties of those that are not.
This would be simplified quite a bit if we kept metadata on attachments.
We can define an inferred node as "successful" if it subsequently becomes an attachment point. It would be interesting to look at how many nodes we infer are successful, and to examine the properties of those that are not.
This would be simplified quite a bit if we kept metadata on attachments.