Version 0.7:
  - am.score() function implements wide-range of standard association measures as well as user-defined formulae

Version 0.6-1:
  - add Lockwords measure to keyness() function
  - keyness() and some other functions now explicitly convert integer vectors to double() in order to avoid numerical overflow and other errors
  - also addresses problem that data.table reads large integer counts as bit64::integer64, which wreak havoc on some built-in maths and stats functions (because they masquerade as doubles); kudos to Mark McGlashan for the bug report with repro code

Version 0.6:
  - keyness() function implements best-practice keyness measures according to Evert (2022)
  - some documentation improvements
  - package tests have been added

Version 0.5-2:
  - now requires R >= 3.5.0
  - add Bonferroni correction to prop.cint() and improve documentation
  - improve documentation for central two-sided p-values computed by binom.pval() and fisher.pval()
  - stricter checks that vector arguments are of the same length

Version 0.5-1:
  - maintenance release to comply with maintainer information format required by CRAN
  - fix incorrect gender of package author and maintainer
  - added a few utility functions: qw(), rowVector(), colVector(), corpora.palette(), alpha.col()
  - minor documentation fixes and updated URLs

Version 0.5:
  - consolidated and revised example data sets included in the package
  - various small, but convenient utility functions added 
  - maintainer information updated

Version 0.4-9:
  - transitory version in which data sets and utility functions for the SIGIL course had been moved into a separate package
  - since the new SIGIL package wasn't accepted by CRAN, the data sets have been moved back into corpora
  - intended re-design of corpora package was cancelled
  - in future, it will be used to collect miscellaneous utility functions for analyzing corpus frequency data

Version 0.4-3:
  - interim release to ensure compatiblity with stricter CRAN checks
  - added data set with Biber register features for all BNC texts (from Gasthaus 2007)
  - some minor corrections

Version 0.4-1:
  - large simulated census data set for examples and illustrations in the SIGIL course
  - simulated type-token statistics from English Wikipedia (based on Wackypedia corpus)
  - convenience function for random samples of rows from a data frame (sample.df)

Version 0.4-0:
  - re-launch of the corpora package on R-Forge
  - first version 0.4-0 has only minor changes over previous release 0.3-2
