1.1

-   implemented penalty modification factors and penalty change distribution
    via a connection matrix
-   implemented estimation of models for competing risks

1.0-1

-   implemented data adaptive rule for default penalty value
-   fixed bug where output of the selected covariate would print the
    wrong name in presence of unpenalized covariates
-   Boosting now starts a step 0, i.e., also the model before updating
    any of the coefficients of the penalized covariates is considered.
    However, the unpenalized covariates will already have non-zero
    values in boosting step 0.
    This change breaks code that relies on the size of elements
    "coefficients", "linear.predictors", or "Lambda" of CoxBoost objects
-   implements parallel evaluation of cross-validation folds,
    via package 'snowfall'
-   speed improvements by replacing 'apply' and 'rbind' , most noticeably 
    for a large number of observations with a small number of covariates


1.0

* initial public release