Version 4.2-7

i) gamlss

1. gamlssML(): now allows the fitting binomial data (sorry it never checked before) and the use of formula in the specification of the model (e.g, y~1) to be consistent with gamlss(). Note that explanatory variables will be ignored if used with gamlssML().
2.  .gamlss.multin.list is now on NAMESPACE
3.  the functions vcov.gamlss() and summary.gamlss() have now an extra argument Hessian.fun with two alternatives “R” and “PB” in order to calculate the Hessian. The “R” option uses the R function optimHess() while the “PB” uses a local function based a function of Pinheiro and Bates in nlme package
4.  the function plot2way() is introduced as a way of plotting a two way interaction between two
   categorical variables (factors).
5. the functions vcov.gamlss() and gen.likelihood() now works even if NA are in the coefficients
   of the beta the paraneters
6.  The functions loglogSurv1(), loglogSurv2(), loglogSurv3(), loglogSurv() and logSurv() are
   introduced as a way of exploring the tail behaviour of a distribution.
7. AIC(), GAIC() and extractAIC() have a new argument “c” which if k=2 and c=TRUE gives the corrected AIC that is, AICc (suggested by Mario Alvarado).

ii) gamlss.dist

1. The use of the distributions BCCGo(), BCPEo() and BCTo() (which are identical to BCCG, BCPE
   and BCT respectively apart from the fact that they using log link for mu) had some side effects
   because the equivalent d, p, q and r functions did not exist. In this version to avoid the
   problem we have added the functions dBCCGo(), pBCCGo(), qBCCGo(), rBCCGo(), dBCPEo(),
   pBCPEo(), qBCPEo() and rBCPEo(), dBCTo(), pBCTo(), qBCTo() and rBCTo(). Those functions
   are identical to d, p, q, and r functions of BCCG(), BCPE() and BCT().

iii) gamlss.tr

1. functions fitTail() and fitTailAll() are introduced for fitting truncated distributions to
   the tails and creating Hill type of plots respectively.

iv)  gamlss.demo

1.  The distributions SN1, SN2, LOGITNO, LOGNO2, TF2 and SST are added to the demos