                        Changes in version 0.2.0                        

Scientific corrections

  - Corrected the Robson--Whitlock point estimator and separated it from
    its one-sided upper confidence bound.
  - Added the unbiased Strauss--Sadler endpoint estimate instead of
    reporting its confidence bound as the point estimate.
  - Corrected the OLE covariance matrix and added a configurable k for
    the most recent sighting events.
  - Corrected conditioning on the observation origin in the Solow
    (1993), declining-rate Solow (1993), and Burgman et al. (1995)
    procedures.
  - Replaced the unsupported former solow2005() calculation with the
    Weibull extreme-value p-value from equation 16 of Solow (2005).
  - Clarified that frequentist outputs are p-values rather than
    posterior probabilities of persistence.
  - Added regression tests based on published examples for the dodo,
    Caribbean monk seal, and black-footed ferret.

New methods

  - Added solow_roberts2003() for the nonparametric two-endpoint
    persistence test.
  - Added jaric2010() for average-interval and sighting-trend inference.
  - Added jaric_roberts2014() for records with individual sighting
    reliabilities.
