Discuss right-truncation procedure for vintaged runs that predate an inferred PMF

[dylanhmorris]

For report dates far enough in the past, there will not be a PMF with an appropriate reference date, and since we were planning not to attempt right-truncuation modeling for the most recent data, I thought simplest might be to use a fixed PMF. But we could of course always use the earliest PMF for report_dates before the first PMF reference_date (or fit PMFs for the earlier period based on the vintages! but we shouldn't try to do that IMO)

Originally posted by @dylanhmorris in #87 (comment)