- bootstrapping models implemented
- improved handling of model classes – included T, t0, N, and X0 in object def
- implemented plotting for model bootstrap
- difficulties with changePt model fit remain. fixed one source of trouble – no longer passes zero-length data to OU.lik() when t_shift is outside the time interval.
- Still giving errors on coercing lists into doubles for the saddle_node bifurcation simulated data set. –
**fixed**, was getting t_shift as a vector from time(X).

- Still giving errors on coercing lists into doubles for the saddle_node bifurcation simulated data set. –

## Code Tricks

Duncan pts out: the S3 mechanism is

```
custom =
function(x, y, ...)
UseMethod("custom")
```

Then that will find and use custom.foo, etc. based on the class of the value of x. You can have it dispatch on a different parameter, but that is not very common. And to dispatch on the types/classes of multiple parameters, e.g. x is a numeric vector, y is a matrix, you would need to use S4.

See the setGeneric() function in R for S4 method (also compatible with S3 classes).