Some nodes will do more work than others and their outputs are unlikely to change. The outputs of those nodes should be cached (and persisted? ) . Its the job of the engine to figure out which nodes are likely to be cached. But perhaps the analyst already knows and we should let them label it.
To be clear, this is a pre-optimization at this point and a low priority
Some nodes will do more work than others and their outputs are unlikely to change. The outputs of those nodes should be cached (and persisted? ) . Its the job of the engine to figure out which nodes are likely to be cached. But perhaps the analyst already knows and we should let them label it.
To be clear, this is a pre-optimization at this point and a low priority