Open ajvincent opened 1 year ago
I really need someone to check my math above.
In particular, I'm imagining revoking one object graph of entirely original objects, no proxies. That's going to be the worst-case scenario in a more-than-two object graphs situation.
I've been thinking about drawing up a table for the repository main page to illustrate what the performance benefit would (or could) be here, in both time and space complexity.
Given m object graphs and n original objects:
The exponential time complexities are scary at first, but really, not that bad. Because the number of object graphs, m is almost always tiny compared to the number of original objects, n.