1.2.5. Re-Fixed bug in MetropolisUpdate
1.2.4. Fixed bug in MetropolisUpdate and changed move for sigma.2
1.2.3. Fixed bug in Birth move (now the prior of sigma2 behaves as 
       documented).
1.2.2. Added vignette in /inst/doc/
       Bug in print.summary.RJaCGH.Chrom
       Typo in chainsSelect.Rd
       Added method for RJaCGH.Chrom in chainsSelect() and bug fixing.
       Explanation of the dimension of k, means, variances and parameters 
       of the transition funcionts in RJaCGH.Rd
       Removed restriction for max.dist to be higher than all probe 
       distances.
1.2.1: Bug in plot.pREC_S (reordering of the array names)
       Now one can choose the method of clustering in plot.pREC_S.
       (default ward method)
       Bug in pREC_S (nor deleting tmp files)
1.2.0: Homocedastic model in RJaCGH
       Change in priors: uniform priors for sigma     
       pREC_A and pREC_S
       Arbitrary names and number of Chromosomes
       Probe names
       Start/End positions
       New get.jump()
       New summary
       Bug in labellling of states
       New vignette
1.1.1: Bug in get jump.
1.1.0: New args to RJaCGH added: max.distance, normal.reference and 
       normal.ref.percentile
1.0.2: Bug Fixing:
       when model=genome there had to be 23 chromosomes
       strange bug with auto.label
       (Thanks to Al Ivens)
1.0,1: Bug fixing: 
       labels in legend of genome.plot
       plot.RJaCGH.array with an even number of chromosomes
       positions must be integers (bases, not Kb).
       (Thanks to Joao Fadista).
1.0.0: Probabilistic minimal commmon regions.
      Plot for the genome (genome.plot).
0.5.: Bug fix in summary.RJaCGH (thanks to Douglas Grove).
0.5:  Automatic labelling of states
0.4-3: Functions can take distance, not just position
       Added preprint to inst/doc
0.4-2: Now collapseChain recovers loglik and probs within model. Changed
       slightly prior for mu.
0.4-1: Fixed bug in plots that smooth arbitrarily too much in density plots
0.4:   Fixed Windows crash (order of evaluation issue in Combine and Split).
0.3-1: Use Calloc instead of R_alloc to try drive down memory usage.
       Played with interrupts.
0.3:   nnhl.cpp re-written as C (instead of C++) to try prevent crash under 
       Windows.
0.2:   Real Posterior probabilities of the sequence of hidden states via viterbi
