Better support for time series history output

[billsacks]

We have received a number of requests for improving the support for time series history output. The two most likely approaches are:

(1) Support in PIO for converting all output to time series output at runtime

(2) Better support for the python tools that convert time slice to time series output – particularly, better portability

However, Gary Strand points out some possible downsides with (1):

Regarding output in timeseries format, the downsides include:
(a) I've seen runs with over 1,000 CAM output fields. Can 1,000+ files be written to simultaneously?
(b) What about model crashes and other causes of run failure? How does the model handle ragged timeseries output, not all of which will even be the same length in time? Stitching timeseries files from runs with overlaps and gaps is a major pain.

[billsacks]

This doesn't seem to be high enough priority that it will happen anytime soon. I'm going to close it as a wontfix; we can always reopen it later if it becomes wanted.

[billsacks]

Reopening to close this properly as incomplete for more clarity.