I have a graph algorithm I don't need to run from scratch when invoking a function because I've been tracking the necessary information in advance. The opposite of lazy evaluation, or "reversed memoization", so to say.
Is there a name for this way of processing data?
I don’t think there’s a standardized term for this, but it seems appropriate to say that you’ve been computing the information incrementally as more data arrive.
For context, there’s a notion of incremental algorithms, which are algorithms that maintain information about a dynamically updated set of values that can change as new data values are added. For example, the “incremental minimum spanning tree problem” is to maintain an MST for a graph as new edges are added in. What you’re describing sounds like it’s in the same ballpark.
