Version 0.1

Version 0.2
  Bug suggested by Bjarne Hansen fixed 7/20/06.  Bug involved augmenting lhs
  samples with one parameter.  Example a <- randomLHS(4, 1); augmentLHS(a, 2).
  RUnit test added to check this bug.  All RUnit tests satisfactory.

Version 0.3
  Another bug reported by Bjarne Hansen on 7/23/06, and fixed on 10/21/06.
  The bug involved augmenting a hypercube with one point.  RUnit test added
  to check this bug.  All RUnit tests satisfactory.  Also added documentation
  consisting of an augmentation example.  Added a lhs package help page.

Version 0.4
  Changed the license to GPL >= 2 according to a Kurt Hornik email

Version 0.5
  Change output filenames to be portable

Version 0.6
  Added a new option to randomLHS to allow for similar lhs's when
  the seed is set and columns are added

Version 0.7
  Removed test directories to fix them for the proper package structure.

Version 0.8
  Refactored the underlying C code into C++ to add range checks
  for internal arrays.  Corrected a bug suggested by XXXX on DATE.
  Corrected the bug suggested by Prof Ripley on DATE with the
  range checking.

Version 0.9
  Removed non-portable code introduced in Version 0.8

Version 0.10
  Changed static template method definitions to be included in the
  definition of the utilityLHS class, instead of in the header below the class.

Version 0.11
  Fixed a bug in the geneticLHS code and added Maximin to the optimization criteria.
  Fixed a bug caused by a change in R2.3.3 that requires a matrix dimnames to be a list.  Added an importFrom to
  the namespace.

Version 0.12
  Fixed a bug in the Description file.

Version 0.13
  Numerical accuracy checks on the Solaris 10 systems was failing.  Unable to reproduce this error easily,
  but it is limited to the numerical accuracy tests.  Eliminated the numerical tests for Solaris/Sun/Sparc.

Version 0.14
  Fixed a bug suggested by Roland Lowe on 8/4/2016.

Version 0.15
  Added registration of native routines.

Version 0.16
  Removed the file output of test files which were causing errors on CRAN.
  changed the version dependency to >= 3.3.0.  (>=3.4.0 caused errors on CRAN)

Verion 1.0
  Major revision.
  Changed all underlying C code to C++ using Rcpp.
  Added orthogonal array latin hypercube capability

Version 1.0.1
  Update to fix a memory leak noticed on CRAN servers when checking examples with valgrind

Version 1.0.2
  Added references to the Orthogonal array functions
  Fixed typos
  Changed the way the Rcpp::RNGScope object is destructed based on the debugging efforts of @mb706

Version 1.1.0
  Bug reports from Dr. Ulrike Groemping (https://github.com/bertcarnell/lhs/issues/26, https://github.com/bertcarnell/lhs/issues/25)
  Fixed underlying C++ code for addelkempn orthogonal array algorithm.
  Fixed underlying C++ for bosebushl orthogonal array algorithm.
  Added regression tests

Version 1.1.1
  Corrected memory leak and read out-of-bounds

Version 1.1.2
  Exposed functions related to Galois fields from the oa C++ library

Version 1.1.3
  Corrected compilation flag issue on Solaris

Version 1.1.4
  Correct error in logical statement with length > 1
  
Version 1.1.5
  std::iterator, std::binary_function, and std::unary_funcation were removed since they are deprecated in c++17 and later