There are use cases (e.g., checkpointing) where it might be useful to have a particular handler in the evaluator system not evaluate on the first iteration. As a positive, this change would reduce startup time and disk usage. As a negative, this change would lose the automatic recording of the initial conditions (but would be user controlled and default to false)
There are use cases (e.g., checkpointing) where it might be useful to have a particular handler in the evaluator system not evaluate on the first iteration. As a positive, this change would reduce startup time and disk usage. As a negative, this change would lose the automatic recording of the initial conditions (but would be user controlled and default to false)
Usage might go something like this:
with a rough sketch of the implementation looking something like this: