Changes

Sébastien Villemot · 92954614
--- a/MATLAB-Versions.md
+++ b/MATLAB-Versions.md
 This page lists changes introduced in MATLAB versions that are the most
 relevant to Dynare development. This information mostly comes from the [MATLAB
 Release Notes](https://fr.mathworks.com/help/matlab/release-notes.html).
 # MATLAB 24.2 (R2024b)
 - New [package manager](https://www.mathworks.com/help/matlab/matlab_prog/use-the-matlab-package-manager-to-organize-and-distribute-code.html)
 - New `isapprox` function for determining approximate equality of elements in two arrays
 - New `RegularizationFactor` name-value argument to the `lsqminnorm` function,
  for Tikhonov regularization to the least-squares solution
 - Many I/O functions can now directly read data online using an HTTP/HTTPS URL,
  notably:
  + `isfile`, `dir`, `copyfile`, and `exist`
  + `fopen`, `fread`, `fgets`, `fgetl`, `feof`, `fclose`, `ferror`, `fseek`, `ftell`, `fscanf`,
    `frewind`, and `fileread`
  + `h5disp`, `h5info`, `h5read`, `h5readatt`, `H5F.open`, and `H5F.is_hdf5`
 - Dictionaries in JSON files can now be read and written using the
  `readdictionary` and `writedictionary` functions
 - Symbolic links support
  + new `createSymbolicLink` and `isSymbolicLink` functions
  + new `ResolveSymbolicLinks` name-value argument with the `delete` and
    `rmdir` functions
  + new `CopyLinkBehavior` name-value argument with the `copyfile` function
  + new `MoveLinkBehavior` name-value argument with the `movefile` function
 - When reading spreadsheet files, the behavior for merged cells can now be
  controlled using new `MergedCellColumnRule` and `MergedCellRowRule`
  name-value arguments of `readtable`, `readtimetable`, `readmatrix`,
  `readcell`, `readvars` and `detectImportOptions`
 - The `zip` and `unzip` functions now support password-encrypted files
 - The `sftp` function now accepts a timeout value and supports passphrases and
  certificates
 - The `summary` function now works on array data, including `numeric`, `logical`,
  `datetime`, `duration`, and `calendarDuration` types
 - The `isbetween` function now works on numeric, nonnumeric, or tabular input data
 - The `rmmissing` and `rmoutliers` functions now accept an argument for
  defining the location of missing entries or outliers using array or table
 - The `table` and `timetable` data types now have a `VariableTypes` property
  for specifying the data types of their variables
 - New `violinplot` function for creating violin plots for grouped numeric data
 - New `compassplot` function to plot one or more sets of arrows originating
  from the origin in polar coordinates
 - New `constantplane` function to create infinite planes for highlighting
  regions of interest in 3-D plots
 - New `IconColumnWidth` property of a `Legend` object for controlling the width
  of the legend icons and the width of the legend box
 - New `Labels` and `LabelLocation` properties of a `Bar` object for adding
  labels to ends of bars in bar charts
 - The `tiledlayout` can now be called without dimensions or tile arrangement
  argument
 - The MAT-file version can now be specified with the `savefig` function
 # MATLAB 24.1 (R2024a)
 - Local functions can now be defined anywhere in scripts
 - Packages (“+” folders) are now called “namespaces” (this is only a change in
  terminology, not in functionality)
 - New `createArray` function for generating arrays of specified class and size
 - New `islocalmin2` and `islocalmax2` for finding local extrema in 2 dimensions
 - New `clip` function for clipping values to a specific range
 - New `Weights` parameter to the `mean` and `median` functions
 - The `iqr` function can now return the first and third quartiles of the input
  data by using a 2nd output argument
 - New `openedFiles` function that returns the file identifiers of all open
  files as a row vector
 - New `pagelsqminnorm` that calculate minimum-norm least-squares solutions to
  systems of linear equations in N-D arrays
 - New `pagepinv` function that calculates Moore-Penrose pseudoinverses of pages
  of N-D array
 - The `spdiags` function now supports implicit expansion for the diagonal
  elements argument Bin. You can specify the diagonal elements as a matrix,
  vector, or scalar, and the function expands the values if the argument is
  specified as a vector or scalar.
 - New `fpolarplot` function for plotting mathematical expressions in polar
  coordinates
 - The Python Interface now converts between MATLAB and Python dictionaries, and
  between MATLAB tables and Python Pandas DataFrames
 # MATLAB 23.2 (R2023b)
 - New dictionary functions: `configureDictionary`, `lookup`, `insert`, `remove`
 - New `mustBeSparse` validation function
 - New `resize`, `paddata` and `trimdata` functions for changing the size of
  data by adding or removing elements
 - New `smoothdata2` function for smoothing noisy data in two dimensions
 - New `kde` function for estimating a probability function by using kernel density estimation for univariate data
 - New syntax for the `rng` and `tallrng` functions where only the algorithm is
  specified (but not the seed, which then defaults to zero)
 - The `readstruct` and `writestruct` functions can now manipulate JSON files
 - New `cmdsep` function which returns the (shell) command separator for the
  current platform
 - New `expmv` function for computing a matrix exponential multiplied by a
  vector
 - New `piechart` and `donutchart` functions for creating configurable,
  interactive pie and donut charts
 - New `Direction` property of the `legend` function for controlling the order
  of legend entries
 - New `xscale`, `yscale` and `zscale` functions for quickly change between
  logarithmic and linear axes scales after plotting
 - New `xsecondarylabel`, `ysecondarylabel`, and `zsecondarylabel` functions for
  creating, deleting, or modifying secondary axis labels
 - New `xregion` and `yregion` functions for creating unbounded regions and
  specifying multiple regions as matrix
 - New `gitrepo` object for manipulating git repositories
 # MATLAB 9.14 (R2023a)
 - Warning about indexing with no subscripts (`MATLAB:subscripting:noSubscriptsSpecified`) has been removed
 - New `combinations` function for generating all element combinations of arrays
 - `head` and `tail` functions now work on arrays
 - New `pageeig` function for calculating eigenvalues and eigenvectors of the
  pages of N-D arrays. In this context, the N-D array is treated as a container
  for several 2-D matrices.
 - The `randi` function can now create random logical arrays
 - Many functions for descriptive statistics and arithmetic (e.g. `max`, `min`, `mean`, `std`, `var`, `median`, `sum`, `prod`, `cummax`, `cummin`, `cumsum`, `cumprod`) now accept the `omitmissing` and `includemissing` options (the former being the default) (the older `omitnan` and `includenan` options are still supported)
 # MATLAB 9.13 (R2022b)
 - New `dictionary` object (hash table)
 - Validation of output arguments in functions and class methods
 - New `mape` and `rmse` functions for calculating forecast errors (mean absolute percentage and root mean-squared)
 - New `isuniform` function for determining if vector is uniformly spaced
 - The `stackedplot` function can now plot variables from multiple input tables or timetables
 - The `detrend` function can now detrend data in a table or timetable.
 - New `pagenorm` function for calculating norms using pages of N-D arrays
 - The `errorbar` function can now plot multiple lines with error bars at once
 - New `tightPosition` function for getting the location and size of axes plotting area
 - New `tilenum` and `tilerowcol` functions for convert between rows and columns and tile numbers in tiled chart layouts
 - New Code Analyzer app (and programmatic interface through `codeAnalyzer`)
 # MATLAB 9.12 (R2022a)
 - New `allfinite`, `anynan` and `anymissing` functions to examine the elements of an input array
 - The `quantile`, `prctile` and `iqr` functions are now part of core MATLAB (they were previously in the Statistics and Machine Learning Toolbox
 - The `rms` (root mean square) function is now part of core MATLAB (it was previously in the Signal Processing toolbox)
 - The `std` and `var` functions now optionally return the mean as second output argument
 - New `writelines` function for writing plain text to a file
 - New `pagemldivide`, `pagemrdivide`, and `pageinv` functions: Solve linear equations and calculate matrix inverses using pages of N-D arrays
 - New `tensorprod` function: Calculate tensor products between two arrays
 - The `round` function has a new `TieBreaker` name-value argument to specify how to break ties. 
 - `null` and `orth` functions have a new option to specify tolerance to treat singular values below a threshold as zero
 - The `norm` function can now calculate Frobenius norm on N-D arrays
 - The `equilibrate` function now has an option with values of `"vector"` or `"matrix"` to specify whether the output arguments are returned as vectors or matrices
 - The `rand`, `randi`, and `randn` functions now support complex input and a `RandStream` object for the `"like"` syntax
 - The `eps`, `flintmax`, `intmax`, `intmin`, `realmax` and `realmin` functions now accept the `"like"` syntax to return scalars based on a prototype object
 - `qr` and `gsvd` have a new `"econ"` option for economy-size decompositions
 - The syntax `R = qr(A)` now always returns `R` as an upper-triangular matrix, regardless of whether `A` is full or sparse (behaviour change)
 - New Code Compatibility Analyzer app
 # MATLAB 9.11 (R2021b)
 - New `parfeval` syntax for parallelizing calls to a function
 - New `trenddecomp` function for trend/cycle decomposition
 - New `ComparisonMethod` option to `min` and `max` functions
 - New `sftp` function for transferring files from/to SSH servers
 - New `outputFormat` option of the `svd` function
 - New `exportgraphics` function to append graphics to an existing PDF file
 - New `pyrun` and `pyrunfile` functions to run Python code
 # MATLAB 9.10 (R2021a)
 - New `name=value` syntax for passing name-value arguments
 - `format` function can now get and set display format
 - `fillmissing` function now accepts a custom method
 - `normalize` is now able to normalize multiple data sets with same parameters
 - XML support in `readtable`, `readtimetable`, `writetable`, `writetimetable`,
  `detectImportOptions`
 - Read over HTTP or HTTPS support for `readtable`, `readtimetable`, `readvars`,
  `readstruct`, `readmatrix`, `readcell`, `readlines`, `detectImportOptions`
  and others
 - New `PrettyPrint` option to `jsonencode` for adding indentation
 # MATLAB 9.9 (R2020b)
- New `pattern` and `extract` functions for string manipulation
+- New `height` and `width` functions (can be used as a replacement for `rows` and `columns` that are present in Octave and that we reimplemented as missing functions; NB: the MATLAB documentation erroneously says that these functions were introduced in R2013b)
- New `mustBe*` functions for input validation
+- New `pattern` and `extract` functions for string manipulation
- New `height` and `width` functions
+- New `mustBe*` functions for input validation
 - New `readstruct` and `writestruct` functions for reading/writing structured
  data in XML files
 - New `readlines` function to read a whole text file as a string array
 - New `pagemtimes` function for batch matrix multiplication
 - New `svdsketch` function for computing SVD factors of low-rank matrix sketch
 - New `matlabRelease` and `isMATLABReleaseOlderThan` functions, meant as
  replacements for `ver` and `verLessThan` respectively
 - C Matrix API `mxCreateString` function now accepts UTF-8 encoded data
 - C MEX API `mexPrintf`, `mexEvalString`, `mexEvalStringWithTrap`,
  `mexErrMsgIdAndTxt` and `mexWarnMsgIdAndTxt` functions now accept UTF-8
  encoded data
 # MATLAB 9.8 (R2020a)
 - The `sparse` function now accepts vectors of integer data type for its `i` and `j` arguments
 - Support for sparse complex arrays
 # MATLAB 9.7 (R2019b)
 - It’s now possible to use the dot-indexing operator on a function call, e.g.
  `createStruct(3).aField`
 - The `size` function now accepts a vector dimension argument to query multiple
  array dimension lengths at a time
 - It’s now possible to restrict the possible values of input arguments to a
  function, using function argument validation
 - Numbers can be specified using hexadecimal or binary literals
 - New `matches` function to determine if two input strings are equal
 - New `namedargs2cell` function to convert structure containing name-value
  pairs to cell array
 - The `error` function now accepts a suggested fix for an uncaught exception
  (should be the first argument, as a `Correction` instance)
 - New `groupfilter` function for filtering data in a table, timetable, or
  matrix by group
 - New `sheetnames` function to get names of worksheets from spreadsheet file
 - New `makima` function for performing modified Akima cubic Hermite
  interpolation
 - New `tiledlayout` function for creating tiled chart layout (now recommended
  over `subplot`)
 # MATLAB 9.6 (R2019a)
 - New `append` function for combining strings
 - The `min` and `max` functions:
  + now return index information corresponding to the minimum and maximum
    values When simultaneously operating on more than one dimension
  + now return the linear indices corresponding to the minimum and maximum values of
    the input array using the `'linear'` option
 - New `readmatrix`, `readvars`, and `readcell` functions for reading tabular
  data as a matrix, variables, or a cell array
 - New `writematrix` and `writecell` functions for writing tabular data from a
  matrix or cell array to a text or spreadsheet file
 - New functions `matchpairs` and `equilibrate` for solving assignment problems
 - New `xcorr` and `xcov` functions for compute cross-correlation and
  cross-covariance (they were previously in the Signal Processing toolbox, now
  they are in MATLAB core)
 - The `detrend` function can now remove piecewise polynomial trends, set
  continuity requirements, and specify sample points
 - It is now possible to provide a suggested fix for an uncaught exception in
  the `MException` class, using the `addCorrection` method and the `Correction` class
 - New `parallelplot` function for visualizing tabular or matrix data with
  multiple columns by using a parallel coordinates plot
 - The `grouptransform` function now operates on arrays
 # MATLAB 9.5 (R2018b)
 - It is now possible to use string arrays where character vectors or cell
  arrays of character vectors were previously accepted; using double quotes
  everywhere is now encouraged
 - New `convertContainedStringsToChars` function to convert string arrays at any
  level of cell array or structure
 - The following functions now accept a vector dimension argument to specify
  multiple operating dimensions at a time, as well as the option `'all'` to
  specify all dimensions of an array: `all`, `any`, `bounds`, `max`, `mean`,
  `median`, `min`, `mode`, `prod`, `std`, `sum`, `var`
 - New `sinpi` and `cospi` functions for computing the sine and cosine of multiples of π
 - `readwrite` and `writetable` are now able to manipulate Excel files on Windows without launching Microsoft Excel
 - New `grouptransform` function for transforming by groups on tables and timetables
 - New `groupsummary` function to perform group summary computations on matrices
 - New `rmoutliers` function for removing outliers in an array, table, or timetable
 # MATLAB 9.4 (R2018a)
 - New interleaved complex ABI for MEX files
 - New C++ MEX interface
 - New `normalize` function for normalizing array, table, and timetable data
 - Possibility of creating complex empty arrays using functions such as `zeros`
  and `ones` with the `'like'` option, or with the `complex([])` syntax
 # MATLAB 9.3 (R2017b)
 - New `isfile` and `isfolder` functions
 - New `decomposition` object to solve linear systems repeatedly with improved
  performance
 - New `lsqminnorm` function to find minimum-norm solution of underdetermined
  linear system
 - New `dissect` function to reorder sparse matrix columns using nested
  dissection ordering
 - New `vecnorm` function to compute vector-wise norms of arrays
 - New values for the `sigma` option of `eigs`
 - The `eigs` and `svds` functions now accept name-value pairs for options
 - The `griddedInterpolant`, `interp1`, `interp2`, `interp3`, and `interpn`
  functions now support the interpolation method `'makima'`, which is the
  modified Akima cubic Hermite interpolation method.
 - New `tallrng` function for controlling random number generation for tall arrays
 - New `isStringScalar` function for determining whether input is a string array
  with one element
 - New `convertStringsToChars` and `convertCharsToStrings` functions to enable
  one’s code to accept all text types as inputs without otherwise altering the
  code
 - The `arrayfun`, `cellfun`, and `structfun` functions can return object arrays
  of any data type, so long as the objects can be concatenated
 - New `geobubble`, `wordcloud` and `binscatter` graphics functions
 - The `bar` function makes it possible to control individual bar colors
 - The `jsonencode` function encodes MATLAB `NaN` and `Inf` values in JSON as
  null, unless asked otherwise with the `'ConvertInfAndNaN'` option set to false. 
 - New `ischange` function to detect abrupt changes in data
 - New `islocalmin` and `islocalmax` functions to detect local minima and maxima in data
 - New `mink` and `maxk` functions for finding the k smallest or largest elements in an array
 # MATLAB 9.2 (R2017a)
 - String arrays can now be constructed using double quotes
 - New `bounds` function for simultaneously determine the smallest and largest
  elements of an array
 - New `'omitnan'` option for the `prod` and `cumprod` functions
 - New `movmad` and `movprod` functions for computing moving median absolute
  deviation and moving product of an array
 - New `missing` function to represent missing data in various types of
  containers and objects
 - New `issortedrows` to determine if matrix or table rows are sorted
 - New `'ComparisonMethod'` and `'MissingPlacement'` options for the `sort` and
  `sortrows` functions
 - New `'monotonic'`, `'strictmonotonic'`, `'strictascend'`, and
  `'strictdescend'` options to the `issorted` function
 - The data type of the output argument for the `compose`, `extractAfter`,
  `extractBefore`, `extractBetween`, `join`, `split`, and `splitlines`
  functions is now the same as the type of input arguments (they would always
  return string arrays in R2016b)
 - New `head` and `tail` functions for returning the top or bottom rows of table
  or timetable
 - New `heatmap` function
 - New `-nocompression` option to the `save` command
 - `jsondecode` now converts JSON null values in numeric arrays to `NaN`
 - New `isoutlier` and `filloutliers` functions for detecting and replacing
  outliers in an array or table
 - New `smoothdata` function for smoothing noisy data in an array or table with
  filtering or local regression
 - New `summary` function for calculating summary statistics and variable
  information in tables and timetables
 # MATLAB 9.1 (R2016b)
 - Adds “automatic broadcasting”, also known as “implicit expansion” (`bsxfun`
  becomes therefore essentially useless)
 - It is now possible to define local functions in scripts
 - New `contains` function
 - New `join`, `split`, `splitlines`, `compose`, `extractAfter`,
  `extractBefore`, `extractBetween` functions
 - New `jsondecode` and `jsonencode` functions
 - Introduces string arrays (with function `string`)
 - New `ismissing`, `standardizeMissing`, `fillmissing`, and `rmmissing`
  functions for finding, filling, and removing missing data from arrays or
  tables
 - The `dir` function now accepts wildcards
 - New `timetable` data container and `timerange` function
 # MATLAB 9.0 (R2016a)
 - No more 32-bit support
 - `writetable` is now able to write Excel files on GNU/Linux and macOS
 # MATLAB 8.6 (R2015b)
 # MATLAB 8.5 (R2015a)
 - New `mxIsScalar` function in MEX interface (but the function was apparently already present in R2014a, though not officially supported nor defined in headers)
 - New `webwrite` function
 - First version to officially support Windows 10, see https://fr.mathworks.com/matlabcentral/answers/223444-is-matlab-compatible-with-windows-10
 - Many functions for descriptive statistics and arithmetic (e.g. `max`, `min`, `mean`, `std`, `var`, `median`, `sum`) now accept the `omitnan` and `includenan` options (the former being the default)
 # MATLAB 8.4 (R2014b)
 - New `webread` function
 # MATLAB 8.3 (R2014a)
 - Support for new Ethernet interface naming under GNU/Linux (older versions are
  difficult to install on modern GNU/Linux systems), see https://fr.mathworks.com/matlabcentral/answers/73938-activating-on-linux-without-eth0
 - New `isdiag`, `isbanded`, `issymmetric`, `ishermitian`, `istril`, `istriu`,
  and `bandwidth` functions for testing matrix structure
 - New `sylvester` function for solving the Sylvester equation
 # MATLAB 8.2 (R2013b)
 - New `readtable` and `writetable` function (the latter not yet capable of writing Excel files on GNU/Linux and macOS, and needing Microsoft Excel installed on Windows)
 # MATLAB 8.1 (R2013a)
 - New option `stable` to `intersect` and other set manipulation functions
 - New `strsplit` and `strjoin` functions
 # MATLAB 8.0 (R2012b)
 - Syntax for abstract classes
 # MATLAB 7.14 (R2012a)
 - `xlsread` now natively supports `.xlsx` files on all platforms (no need to
  have Microsoft Office installed)
 # MATLAB 7.13 (R2011b)
 - New `narginchk` function
 # MATLAB 7.12 (R2011a)
 - New random number generators interface (`rng`, `getGlobalStream` and `setGlobalStream` functions)
 - New `ichol` function
 # MATLAB 7.11 (R2010b)
 - New `isrow`, `iscolumn` and `ismatrix` functions
 # MATLAB 7.10 (R2010a)
 - Ships with MKL 10.2, which fixes a nasty bug in `dgges` (for which a
  workaround has been implemented in the `mjdgges` MEX)