[Discussion] Task Trigger Criteria

Currently all events trigger based on simulation time in some fashion.

Extracting the event trigger from the event and allowing a larger degree of input specification for what triggers an event could provide significant utility.

e.g.
we could trigger an task on a failure of the linear solver or nonlinear solver to converge which dumps matrix / residual information,
or we could monitor a specific geometric or material property and only trigger time history collection when a certain threshold is passed.

I'm sure there are numerous other possible use cases.

This could also add unwanted complexity to problem specification, or make debugging runtime failures a bit more complicated, etc.