MEFA VERSIONS

Version 3.2-9 -- May 19, 2024

    * Replace deprecated vegan::adonis with vegan::adonis2.
    * Maintainer email changed to personal email.

Version 3.2-8 -- Oct 6, 2021

    * Added DOI for JSS article to replace URL.

Version 3.2-7 -- Jan 11, 2016

    * Vignette produced error/warning due to unsafe use of '...'.
      The names could match current or future control arguments such as
      'decreasing'.  A sometimes-encountered unsafe practice is to call
      do.call('order', df_obj) where df_obj might be a data frame.
      Fix: copy df_obj and remove any names in report.mefa.

Version 3.2-6 -- Nov 21, 2015

    * stack.dist gave wrong labels.
    * cleanup to Rd files.


Version 3.2-5 -- December 3, 2013

    * ade4 based cca example removed from vignette and demo,
      ade4 dependency removed as a result.


Version 3.2-4 -- Sept 25, 2013

    * Spell check completed using utils:::aspell.
    * Removed ::: from mefa:::rep.data.frame call,
      mefa4:::Mefa and mefa4:::Xtab to satisfy R 3.0.2 checks.
    * Depends: R (>= 2.14.0) because vignette directory has moved.


Version 3.2-3 -- July 11, 2012

    * Startup message: line break removed.
    * inst/COPYING removed.
    * Vignette moved to vignettes from inst/doc.


Version 3.2-2 -- May 14, 2011

    * mefaCrosstab: xtabs/table class attribute removed
      from return object.
    * packageStartupMessage used in zzz.R .onAttach


Version 3.2-1 -- May 13, 2011

    * Vignette: value of 'strip.white' option is now lowercase.


Version 3.2-0 -- February 7, 2011

    * CRAN release along with first release of mefa4.


Version 3.1-7 -- January 21, 2011

    * zero sample handling bug fixed in stcs
    * coercion methods for S4 classes Xtab and Mefa added
    * if mefa is loaded after mefa4, as.array gives a warning
      because of Matrix overloads base as.matrix and as.array,
      so mefa should be called before mefa4
    * accessors work irrespective of loading sequence
      of mefa and mefa4 (still masked out but it has no
      effect on functionality.
    * segm() returned NULL instead of $xtab when segm was NULL,
      this is fixed.
    * rdrop/cdrop can take mefa orMefa object as argument.
    * Namespace clash caused by xtab, samp, taxa generics fixed
      (S4 objects were accessed wrong when mefa4 was overloaded
      by mefa).
    * Matrix::as.matrix() is used inside cdrop/rdrop when x is
      an S4 Mefa object (col[row]Sums failed - now fixed).
    * as.mefa.list can handle lists of sparse matrices.


Version 3.1-6 -- August 23, 2010

    * image plot for 'dist' objects
    * aoplot method added for abundance-occupancy
      plots as in Digby and Kempton 1987 p. 40.
    * New extractor methods added for mefa objects:
      xtab, samp, taxa, segm


Version 3.1-5 -- June 4, 2010

    * as.data.frame.dist renamed to stack.dist
    * as.mefa.default now return call correctly
    * plot returns histogram if data is continuous
      and barplot if data is discrete

Version 3.1-4 -- November 10, 2009 (closed - r486)

    * .onUnload problem fixed (reported by B D Ripley)


Version 3.1-3 -- October 15, 2009 (closed - r484)

    * CRAN minor release.


Version 3.1-2 -- September 22, 2009 (closed - r478)

    * mefa got namespace.
    * reshape function in reshape is no longer
      available, citations now pointing to cast
      instead (in manual and vignette).


Version 3.1-1 -- July 10, 2009 (closed - r477)

    * dist2vec: deleted (use as.vector instead)
    * fill.na: accepts 1 dimensional vectors.
    * as.data.frame.dist: now returns labels of the
      dist object (dim.names=TRUE).
    * inflate: is now obsolete, rep.data.frame is
      used instead.


Version 3.1-0 -- June 14, 2009 (closed - r407)

    * CRAN major release.


Version 3.0-3 -- June 14, 2009 (closed - r406)

    * distance matrix manipulation tools added (is.dist,
      dist2vec, as.data.frame.dist)
    * report.mefa: dir argument became more transparent
      by using getwd().


Version 3.0-2 -- May 20, 2009 (closed - r381)

    * vec2dist: function to convert a vector to dist.
    * mfill, rdrop, cdrop: new functions to
      calculate matrix fill and drop rows/columns
      from matrices (similar to na.exclude).
    * documentation encoding problems fixed.
    * mat2list: new function converting a matrix
      into list (row or columnwise).


Version 3.0-1 -- May 1, 2009 (closed - r363)

    * citation info: updated in the Rd files.
    * class attribute of mefa objects: the 'list'
      element has been removed in order to leave
      as.mefa.default work properly.
    * as.mefa/as.stcs: redefined as method, together
      with 'default' class, and a bunch of coercion
      methods added (matrix, list, array, data.frame).


Version 3.0-0 -- March 7, 2009 (closed - r320)

    * Version update: to indicate that this is the
      published version.
    * CITATION: points to JSS paper.
    * vignette: changes made by Achim Zeileis to
      conform to the final version of the JSS paper.
    * Suggests: MASS, ade4 and vegan because of the
      vignette.


Version 2.1-3 -- February 17, 2009 (closed - r318)

    * documentation fixes: due to stricter R CMD
      check of R devel.
    * vignette: minor changes to mirror requested
      changes in JSS manuscript
    * citation: new citation of accepted JSS
      manuscript added.


Version 2.1-2 -- January 6, 2009 (closed - r315)

    * suggests and vignette dependencies: ade4 and
      vegan added, due to error reported by Kurt
      Hornik after CRAN submission.
    * label: assigns and returns a label stored as
      a deparsed attribute. Can be used to store
      relevant information on the current data
      set. Note: inheritance is not implemented,
      so by any transformation of the object
      into another one, it is lost.
    * vignette and demo updated: based on changes
      made in the JSS paper due to second revision.
    * as.data.frame: new method to take advantage
      in modeling. It returns a data frame containing
      variables that are to be used in a model
      i.e. glm. Four helper functions are made to
      support this function, but those can be used
      on their own as well (mss, msr, mts, mtr).
    * plot methods: many tiny changes, image method
      now display names, but by default the axes are
      hided and just a box is drawn.


Version 2.1-1 -- December 11, 2008 (closed - r268)

    * melt.mefa: did not return stcs class if segments
      were present -- bug fixed.
    * mefaTable: xtab.fixed=TRUE behaved like FALSE
      when names had exact match, fixed.
    * demo: got a new ending, by a question for opening
      mefa website.


Version 2.1-0 -- November 17, 2008 (closed - r268)

    * minor release: on CRAN containing additions of
      from 2.0-2 and some polishing.
    * basic object structures and methods: will not change
      in future versions. mefa 2.1 is a yardstick in
      package development. Sorry for any inconveniences
      happened because of restructuring the package.


Version 2.0-2 -- November 12, 2008 (closed - r252)

    * onload text: mefa gives date info when loaded.
    * mefaCompare: compares 2 objects of class mefa,
      based on dimnames and numeric values. Sample
      and taxa tables are not compared.
    * demo: added, based on the vignette (also a tests
      directory to check the demo script).
    * vignette: added to inst/doc based on a manuscript
      submitted to JSS. Dependencies are excluded and
      a verbatim code is placed in the vignette instead.
    * DOLINA data sets: reworked. A more simple example
      is given: dol.count, dol.samp, dol.taxa. The previous
      full data set is available at: hdl:1902.1/12060.
    * extract.mefa: nested segment indexing convention
      introduced when extracted.
    * mefaNestedless: makes an object with non-nested
      segments, to be used in melt.
    * as.mefa: if class is mefa, the object can be
      rebuilt with new arguments (e.g. drop.zero,
      nested, segment).
    * mefaCrosstab: xtabs function is now used for
      crosstabulation internally, instead of the
      inflate function. This implementation allows
      non integer values as 'count'. All other
      functions (stcs, melt, aggregate) accept it, too.
      The inflate function is in use for integer counts
      used by stcs when expand is TRUE. Help pages and
      examples modified accordingly.
    * boxplot: new argument 'all' for plotting all
      segments and the separate segments in the same graph.
    * melt.mefa: now accepts mefa object with xtab
      element without dimnames. In this case, rows and
      columns are names by integers. When only melt(x)
      is used and x has original segments, those are
      used to build the stcs object. This can however be
      avoided by melt(as.mefa(x, segment = FALSE)).
      As a result, ordering of dimnames can be different.
    * mefa: the restriction of at least 2 columns for
      samp and taxa tables is no longer necessary.
    * qvector: na.rm and ... arguments added.
    * plot: labels corrected.
    * melt.mefa: takes care of zero.pseudo flags silently,
      and returns call attribute as well of raw=FALSE.
    * print.mefa: returns segment names to avoid confusion
      when all other details are the same.
    * mefaTables: xtab.fixed argument now works properly.
    * stcs: argument zero.pseudo now accepts a two element
      vector for species and segments, respectively.
    * summary.mefa: returns list with different names
      for [1:4] elements due to consistency with
      samples/taxa sides: s.rich, s.abu, t.occ, t.abu,
      and these are used in the print.summary method, too.
      Accordingly, the class has now a call attribute to
      return it by the summary.
    * help pages: lots of minor improvements.
    * R code: annotated with comments.
    * summary.stcs: returns same info as summary.mefa,
      but print.stcs remained the default data frame method.
    * plot, boxplot, image methods for stcs: also added
      for avoiding the use of calls like plot(mefa(x)).


Version 2.0-1 -- November 8, 2008 (closed - r205)

    * extract.mefa: now accepts negative indexing
      for exclusion (not as character), and
      column extraction bug fixed
    * type arg of plot.mefa: "hist" replaced by "bar"


Version 2.0-0 -- September 29, 2008 (closed - r200)

    * The package was extensively rewritten
      (started on September 16, 2008).
      It is more easy and efficient than
      < 2.0 versions. The number of object
      classes reduced. The new class stcs
      is almost equivalent to the old sscount.
      The new class mefa can be used either
      as the old xcount and mefa classes
      (and the xclist and mflist classes
      only provided with the undocumented
      patch file). The old xorder class
      is bow useless, too.
    * New methods are provided for manipulating
      pbjects. The "[.mefa" replaces the old
      exclmf function, the aggregate method
      replaces the old strify function.
    * plot, print and summary methods only
      provided for the class mefa. For the
      stcs class, methods for data frames can
      be used (because an stcs object is
      essentially a data frame with some
      conventions and additional attributes).
    * The melt method is provided to convert
      a mefa object into stcs, or redefine
      new segments.
    * dim and dimnames methods also available
      for mefa objects.
    * the mefalogo function depicts the
      general representation of an object
      of class mefa.


Version 1.1-4 -- September 20, 2008 (closed)

    * inflate: arguments of any() corrected


Version 1.1-3 -- September 15, 2008 (closed)

    * citation entry added


Version 1.1-2 -- June 24, 2008 (closed)

    * several minor fixes in .Rd files
    * xytrend moved to sedar
    * preliminary modifications made according to later
      introduction of 'list' objects
    * report as generic function added for methods
    * as.mefa added
    * startup message is added


Version 1.1-1 -- June 12, 2008 (closed)
                 April 02, 2008 (started)

	* accumulate was rewritten
	* vignette is corrected and updated
	* help pages are corrected and updated


Version 1.1-0 -- March 14, 2008

	* vignette is made available
	  and can be used as reference for the
	  dolina demo
	* add.attrib can now handle mefa objects
	  containing NULL attribute table
	* logscale argument inserted into plot methods
	* cutoff argument inserted into print methods
	* report.mefa bug fixed & help corrected
	* new functions: qdef, mapmf, accumulate,
	  ivgen, xytrend
	* mefa from now depends on other packages
	  (xytrend requires vegan)
	* mefa developement is now on R-Forge


Version 1.0-4 -- February 11, 2008

	* the demo was rewritten
	  (and still under construction...)
	* amf function is provided
	* class definitions are rewritten
	* print and plot S3 methods provided for classes
	* digits argument is placed into sscount
	  and related functions


Version 1.0-3 -- January 4, 2008

	* mefa object can be used in function as.xcount
	* default setting of index argument
	  in add.attrib was changed
	* "SampleReport" option is available
	  in mefadocs function, showing options for
	  Sweave report generation.
	* bug fixed in exclmf
	* bug fixed in dolina demo.
	* warning message is placed in strify code


Version 1.0-2 -- December 18, 2007

	* new function: exclmf
	* default setting of index argument
	  in xorder and check.attrib has changed
	* thus manual was modified inseveral pages
	* There is a mefadocs function,
	  based on J. Oksanen's vegandocs function
	* currently only "changelog" is active
	* A bug is present in report.mefa,
	  a WARNING is placed in the manual
	  will be fixed later


Version 1.0-1 -- December 7, 2007

	* bugs fixed in function stify


Version 1.0 -- November 13, 2007

	* the first public version