# $Id: NEWS 276 2014-05-04 15:50:04Z thothorn $

	0.9-99992 (03.05.2014)

        cleanups by MM

	0.9-99991 (25.04.2014)

        version 0.9-9999 introduced new bug in dmvnorm

        0.9-9999 (17.04.2014)

        dmvnorm (again) returns NaN in case sigma is not decomposable

        0.9-9998 (21.03.2014)

        faster code for dmvnorm by Matteo Fasiolo <mf364 at bath.ac.uk>

        0.9-9997 (17.01.2014)

	T. Miwa fixed a runtime error in miwa.c reported by UB sanitizer

        0.9-9996 (16.09.2013)

        documentation updates/corrections/examples by Marius Hofert
        df = Inf
        rmvt(): argument 'mean' not allowed anymore (prone to misuse)

        pmvnorm(lower=c(-Inf, 0, 0), upper=c(0, Inf, Inf),
                mean=c(0, 0, 0), sigma=S, algorithm = Miwa())
        returned NaN, fixed by Xuefei Mi


        0.9-9995 (29.05.2013)

        update to version 2013-06-29 of mvtdst.f from Alan's website
        (fixed a bug for 2-dim pmvt)

        0.9-9994 (06.12.2012)

        set.seed(29)
        rmvnorm(10, ...)

        produces the same first ten rows as

        set.seed(29)
        rmvnorm(100, ...)

        as suggested by Paul Johnson <pauljohn@ku.edu>. There is a new
        argument pre0.9_9993 for changing back to the `old' output. This
        _DOES NOT_  apply to rmvt.

        0.9-9993 (22.10.2012)

        sigma is called `scale' matrix of {dpq}mvt, thanks to
        Richard Boys <Richard.Boys@ncl.ac.uk> for the hint

        0.9-9992 (19.01.2012)

        qmvt(..., df = 0, ...) gave NaN

        R CMD check works on x64 with GCC 4.6.2 (Debian 4.6.2-11)

        0.9-9991 (10.06.2011)

        tvpack wasn't working on 64bit machines. The reason
        was that the wrapper SUBROUTINEs around the original
        FUNCTIONs didn't return the appropriate double precision
        for unknown reasons. Redefining TVTL and BVTL as FUNCTIONs
        fixed the problem.

        0.9-999 (26.04.2011)

	still problems in fix approx_interval (when !is.null(sigma));
        disable for the time being

        0.9-99 (21.04.2011)

	fix bug in approx_interval spotted by Ravi Varadhan
        <rvaradhan@jhmi.edu>

        0.9-98 (19.04.2011)

	allow ... to pass arguments to rmvnorm in rmvt

        0.9-97 (31.01.2011)

	use check.attributes = FALSE in isSymmetric calls
        (requested by Nick Sabbe <nick.sabbe@ugent.be>)

        0.9-96 (28.01.2011)

        use fixed interval when sigma is specified in qmvt

        0.9-95 (18.11.2010)

        q{mvt,mvnorm} shall always return a list, not a vector
	checks for correlation matrices are less picky now

        0.9-94 (16.11.2010)

        allow for two different noncentral t distributions (via type
        argument)

        add support for one-dimensional quantiles (requested by Jerry Lewis
        <jerry.lewis@biogenidec.com>)

        documentation fixes for problems spotted by
        Jerry Lewis <jerry.lewis@biogenidec.com>

        0.9-93 (30.08.2010, not published)

        interface to Alan's TVPACK algorithms for 2- and 3-d
        probabilities by Bjoern Bornkamp added.

        0.9-92 (06.07.2010)

        update to new mvtdstpack.f (7/10) by Alan. Fixes
        potential bias problem in higher dimension.

        0.9-91 (13.04.2010)

        better search interval for uniroot in qmv{t,norm} speeds up
        quantile estimation; suggestion by Björn Bornkamp
        <bornkamp@statistik.tu-dortmund.de>

        0.9-9 (27.01.2010)

        document ... in pmvt.Rd

        0.9-8 (27.10.2009)

        add citation entry

        0.9-7 (22.05.2009)

        make sure `error' is not NA

        0.9-6 (25.03.2009)

        update Alan's FORTRAN code

        0.9-5 (17.03.2009)

        fix FORTRAN bug spotted by Alex Lenkoski
        <lenkoski@stat.washington.edu>

	0.9-3 (22.12.2008)

	update meta data

	0.9-2 (08.07.2008)

        be a little more liberal (tol = sqrt(.Machine$double.eps))
        when testing for symmetry of covariance matrices
        (and make R CMD CHECK monomvn happy again)


	0.9-1 (02.07.2008)

        better check for covariance matrices, suggested by
        James Rogers  <James.A.Rogers@pfizer.com>


	0.9-0 (1.04.2008)

	add support for the multivariate normal distributions in small
        dimensions by Miwa's method thanks to Tetsuhisa Miwa and Xuefei Mi;
        both have been added as `authors'.

        new argument `algorithm' defaulting to `GenzBretz()' with
        `Miwa()' being the alternative. Those two functions are now used
        to specify hyper parameters such as `abseps'.

        internal function `mvt' is no longer exported.

	0.8-3

        make sure rmvnorm(1, sigma = matrix(0.5, 1, 1)) works
        (reported by Kurt Hornik)

	0.8-2 (10.02.2008)

	rmvnorm() now issues a warning for non-symmetric sigma and uses
        the eigenvalue decomposition as default.

        make gfortran 4.3 happy


	0.8-1 (24.07.2007)

        Orion Poplawski <orion@cora.nwra.com> spotted a meaningless check in
        the regression tests


	0.8-0 (23.07.2007)

	upgrade to 7/7 version of MVTDST (includes better support for
        dimensions > 100). Thanks to Karen Conneely <conneely@umich.edu>
        for motivating the update and for checking the new version.

	rmvnorm() now can also use a Cholesky decomposition to compute
        the root of sigma (thanks to Fabian Scheipl)


	0.7-5 (15.09.2006)

        fix problem reported by valgrind

	0.7-4 (08.09.2006)

        add long requested `dmvt'
        call RNG functions only one time
        make sure unifrnd is double precision

	0.7-3 (23.08.2006)

        make sure pmvnorm(lo=c(-Inf,-Inf), up=c(Inf,Inf), mean=c(0,0) == 0

	0.7-2 (29.08.2005)

        make gfortran happy (a warning about unused variable NF remains)

	0.7-1 (18.11.2004)

	use #!/bin/sh

	0.7-0 (14.10.2004)

	a coding session with Frank produced `qmv{t,norm}'.
        try to check if the support specified by `lower' and `upper' is
        empty (problem spotted by Peter Thomson <peter@statsresearch.co.nz>)
	Alan's fix prevents negative values to be returned.

	some cosmetics

	0.6-8 (03.06.2004)

	EXIT statements are not supported by `f2c', Alan added GOTO
        statements to `MVCHNC'

	0.6-7 (27.05.2004)

	Alan's fix to MVCHNC solves problems with large degree of freedom

	0.6-6 (22.01.2004)

	`La.eigen' is deprecated and `eigen' replaces it in R-1.9.0

	0.6-5 (14.11.2003)

	check if covariance matrix is pd in rmvnorm (by Fritz Leisch)

	0.6-4 (06.10.2003)

	use new base function `cov2cor'

	0.6-3 (21.07.2003)

	Alans changes were restricted to N <= 100, now N <= 1000
        are possible again

	0.6-2 (25.06.2003)

	Alan's recent changes to `mvt.f' make `g77 -pedantic -Wall'
	happy

	0.6-1 (18.06.2003)

	pmvt(..., df = 0, ...) will return normal probabilities for both
        the univariate and multvariate problem

	0.6-0 (17.06.2003)

	Fortran code in `mvt.f' updated to recent version by Alan and Frank.
        This fixes problems with `pmvt' and large degrees of freedom.

	0.5-15 (16.06.2003)

	a note on one-sided probabilities in `pmvt'
	correlation matrices in cats example a little bit nicer

	0.5-14 (06.05.2003)

	the package owns a vignette based on the paper in RNews 1(2)

	0.5-12 (08.05.2003)

	allow df=0 for pmvt

	0.5-11 (29.04.2003)

	package npmc trys to use 'mvt' which is internal: export it anyway

	0.5-10 (23.04.2003)

	mvtnorm is now in a NAMESPACE

	0.5-9 (13.02.2003)

	log argument added to dmvnorm, thanks to
        Jerome Asselin <jerome@hivnet.ubc.ca>

	0.5-8 (21.01.2003)

	fixed bugreport PR#2478: sigma for univariate probabilities

	0.5-7 (27.11.2002)

	use R's random number generator in the FORTRAN code:
	set.seed has now has the desired impact.

	0.5-6 (07.10.2002)

	rmvt added

	0.5-5 (03.07.2002)

	use .Fortran(..., PACKAGE="mvtnorm")

	0.5-4 (09.04.2002)

	correlation matrices for sigma with unequal variances incorrectly
	computed, added `sig2corr' for that propose, tol argument removed,
	fix by Alan to mvt.f

	0.5-2 (22.03.2002)

	Frank added `tol' argument to MVTDST, now in mvtnorm

	0.5-1 (24.01.2002)

	pmvt(0,1) works now

	0.5-0 (10.12.2001)

	release for R-1.4.0

	0.4-4 (06.12.2001)

	bugfix

	0.4-3 (05.12.2001)

	the length of lower, upper and mean (delta) is now recycled to the
	length of the largest, i.e. it is possible to say
	pmvnorm(lower=-Inf, upper=1, mean=rep(1,10), corr=diag(10))

	0.4-2 (04.12.2001)

	several typos, man-pages improved

	0.4-1 (04.12.2001)

	interface changed: sigma (covariance matrix) can be specified as
                           well
	{rd}mvnorm added from package e1071 (thanks to Fritz!)
