Thanks Bob.
Sure. I work with monotone i-splines to model the cumulative baseline hazard for a survival model; since the cumulative baseline hazard is minus the log of the survival function, which in turn is non-increasing, the log cumulative baseline hazard needs to be non-decreasing. Now, the \gamma are used to find a linear combination of the non-decreasing i-ispline basis functions evaluated at the individual survival (or censored) times.
Moreover the likelihood has not only a term that depends on the log cumulative hazard but also a term that is proportional to the log of the derivative of the log of the cumulative hazard function with respect to time.
For the model, see here.