Version 1.0-3
    o fixed a couple of printing bugs

Version 1.0-2
    o Correct PRI scores for relation "necessity" and neg.out = TRUE
    o New argument "neg.out" for both truthTable() and eqmcc() functions

Version 1.0-1
    o Fixed a small bug that prevented the printing on the screen a solution with only one term.

Version 1.0-0
    o The QCA package is now ready for fuzzy sets (thanks to Alrik Thiem's efforts)
    o eqmcc() and truthTable() are extensively reworked, for compatibility with fuzzy sets
    o When fed with a truth table object (of class "tt"), the (e)qmcc() function automatically
      determines the causal conditions and the outcome
    o New function pof(), which computes parameters of fit for a variety of input options
    o New function findTh(), that finds thresholds for crisp sets calibrations
    o New function calibrate(), that recodes the initial data into either crisp or fuzzy sets 
    o New function superSubset() for analysing necessity and sufficiency
    o New argument "rowdom" to function eqmcc(), which allows maximum flexibility for solving
      the PI chart. The user is now able to choose whether to apply row dominance or not
    o Truth tables dropped columns "freq0" and "freq1", and gained columns "incl" and "PRI"
    o Truth tables can be sorted
    o The truthTable() function now returns a raw truth table with numerical columns, plus a
      new component with the recoded data if fuzzy sets are used
    o prettyTable() and createString() functions are now internal
    o Fixed a very old bug in qmcc() related to the decision of which solution is more parsimonious
      (thanks to Michael Baumgartner for pointing that)
    o findPrimes() is now findSubsets()
    o New printing methods for PI charts - objects of class "pic", parameters of fit - class "pof",
      superSubset() results - class "ss", allExpressions() - class"aE" and factorization results - class "fctr"
    o Function qmcc() has been obsoleted
    o Numerous other bug fixes and improvements

Version 0.6-5
    o qmcc() and eqmcc() now additionally return invisibly the PIchart and the initial combinations

Version 0.6-4
    o The functions qmcc() and eqmcc() now return an invisible list with three components:
      - the minimized PIs
      - the number of k PIs necessary to cover all initial causal combinations in the PI chart
      - the final solution(s)
    o Removed redundant code from verify.data()
    o Changed the error message on failure to solve the PI chart (thanks to Juha Järvi)

Version 0.6-3
    o Fixed a bug in factorize(), the argument "use.letters" was replaced
      by "splitmethod", a matrix unintentionally dropped dimensions

Version 0.6-2
    o Modified version of factorize(), the argument "use.letters" was replaced
      by "splitmethod"
    o Improved documentation and citations
    o Improved version of findPrimes(), eliminating empty prime implicants (line
      1 in the 3^k matrix)
    o Fixed a bug in findSubsets(), did not calculate correctly for the case of
      multi-values

Version 0.6-1
    o Minor improvements to findPrimes() and findSubsets(), their first argument
      "noflevels" now must be specifically declared in the higher dimension matrix
    o Fixed a small bug in eqmcc() that prevented correct multi-value minimization
      when excluding the remainders (thanks to Damien Bol for pointing that)
    o Argument "uplow" in eqmcc() is now TRUE by default

Version 0.6-0
    o eqmcc() now accepts multi-valued data!
    o eqmcc() gained a new argument "uplow" (FALSE by default), to print the
      solution in upper/lower letters if at all possible
    o Improved version of findPrimes(), to accommodate multi-valued data
    o Improved version of createString(), to accommodate multi-valued data
    o Improved version of writePrimeimp(), to accommodate multi-valued data
    o Improved version of the internal function writePrimeimp(), to accommodate
      multi-valued data
    o Modified version of demoChart(), the argument "use.letters" was replaced
      by "splitmethod"
    o prettyString() is completely changed, using the native strwrap() and it is
      now an internal function.
    o qmcc() and eqmcc() were changed accordingly, using the new prettyString()
    o New toy dataset "MV" to test multi-valued data

Version 0.5-4
    o Eliminated the error message if there is only one causal combination to
      explain, it seems this is more informative than a logically equivalent
      error message (thanks to Burt Monroe for pointing that)

Version 0.5-3
    o Fixed a minor bug in truthTable(), now it correctly arranges the columns
      corresponding to the outcome values
    o truthTable() now allows for multi-values, both in the causal conditions
      and in the outcome

Version 0.5-2
    o Fixed a major bug in createMatrix() regarding multi-value data
    o Trapped an error when the PI chart is too large, in function solveChart()

Version 0.5-1
    o Updated and cleaned most function descriptions
    o Functions is.tt() and print.tt() are now user accessible
    o Fixed a small bug in rowDominance2(), a matrix unintentionally dropped
      dimensions
    o Fixed a small bug in eqmcc() for complete truth tables
      (thanks to Ronggui Huang for pointing that)

Version 0.5-0
    o truthTable() now returns an object of class "tt"
    o New internal functions is.tt() and print.tt()
    o qmcc() and eqmcc() now accept a truth table as well as raw data

Version 0.4-9
    o Functions findPrimes() and findSubsets() are now at user level
    o Improved version of factorize(), the former did not detect some
      combinations

Version 0.4-8
    o New internal function findPrimes()

Version 0.4-7
    o Fixed a minor bug in rowDominance2()
    o Improved, visibly faster function eqmcc()
    o Minor improvement to the truthTable() function
    o New internal function findSubsets()

Version 0.4-6
    o Reworked version of rowDominance2()
    o Function shortqca() is now eqmcc(), faster by a factor of at least 2

Version 0.4-5
    o New experimental function shortqca(),  many times faster than qmcc() for a
      large of causal conditions (over 11), with a _very_ low memory consumption;
      it can easily work with 16 or 17 causal conditions. It is only experimental
      but it looks very promising, especially for multi-valued data.

Version 0.4-3
    o Fixed a bug in qmcc(), a matrix unintentionally dropped dimensions upon
      subsetting

Version 0.4-2
    o Fixed a bug in qmcc(), the data was not correctly subsetted.
    o Fixed a bug in qmcc(), involving the argument "show.cases"
    o Added argument "use.letters" to function demoChart()
    o Fixed a minor bug in verify.truthTable()

Version 0.4-1
    o Added a namespace for the package
    o Completely rewritten version of base3rows(), now much faster
    o Reworked version of createMatrix(), now much faster for the same number of
      values in each condition
    o Added option "diffmat" to the main qmcc() function; users can choose to
      generate the differences matrix or not: speed or low memory consumption.

Version 0.4-0
    o Reworked version of qmcc() with a substantially lower memory consumption;
      it can now work with up to 15 conditions in 1GB of RAM
    o New function base3rows()
    o Fixed minor bug in writeSolution()

Version 0.3-5
    o Reworked version of qmcc() which does no longer posess the tt argument
      It is faster, working reasonably with up to 13 conditions.
    o Reworked version of truthTable()

Version 0.3-1
    o New function rowDominance() to ease the solving of large prime implicants
      charts

Version 0.3-0
    o New function getRow() which returnes the combination of levels for
      a specific row number
    o Fixed a minor bug overseen by the changes in version 0.2-6
    o Function createMatrix() now accepts multiple values (not just 0 and 1)
    o qmcc() is now 20 times faster and it can reasonably work with up to 11
      conditions
    o completely redesigned function createChart(), which is now internal
    o old function createChart() is now called demoChart()

Version 0.2-6
    o QCA does not depend on package "cluster" anymore as qmcc() is now based on
      dist() from the standard stats package, instead of daisy()

Version 0.2-5
    o Minor documentation fixes
    o Almost completely rewriten prettyTable() and minor changes to truthTable()
    o QCA does not depend on package "combinat" anymore, as function combn() is
      now part of the "utils" package (loaded by default)
    o New function sortMatrix() in the internal functions
    o Argument verifications are now part of internal functions
    o New function readTosmana() for reading Tosmana's XML data files'

Version 0.2-2
    o Completely redesigned function allExpr() - now allExpressions()
      (way faster than the former)
    o Set up the internal functions; qmcc now looks less cluttered
    o New user level function solveChart() - pulled out of the main qmcc()

Version 0.2-1
    o New function createMatrix() which creates the base matrix for the truth
      table
    o New function base3rows()
    o Fixed minor bug in writeSolution()

Version 0.2-0
    o New function factorize() which finds all common factors of literals in a
      QCA solution
