NEWS

NEWS about R package optimx

=========== ONGOING CONCERNS =============
CURRENT BUGS or ISSUES
      o summary.opm should display solution being C(onstrained) or U(nconstrained) and 
	    parameter status like proptimr(). C and U are attributes of "value". 
      o Put counters in efn, egr, ehess -- at least as a commented template using optsp
        environment. These are NOT output by all methods.
      o Look at masks for all methods using bdmsk elements 0 -- this would be
	    in optimr(). Could it be effected via the gradient??
      o Can we report "known" answers for different test functions?
      o In vignettes, use code=readLines() to incorporate examples rather than inline?
        This allows for easier testing of the examples separately.
      o We will NOT add ipoptr until non-standard install is fixed up.

TODOS ideas:
   o    demo() should be cleaner
   o    On failure of method, what should be returned? Could current optimr result in
        such cases be improved?

   Rtnmin
   o    The routines are a conversion of Stephen Nash tn/tnbc to R. 
	As such, they are not necessarily efficient.
   o    Simplification and streamlining of code to mesh better with other R optimization 
        tools. Also put maxit, maxfun in control list somehow. Setup trace as integer 
        to work properly.
   (These NOT checked as at 20230619)
   o Rtnmin has same issue as Rvmmin resetting bounds bdmsk ??
   o 20171007 -- make lmqnbc printout conditional on trace??
   o ??2016-09-22 account for null gradient 
   o Rtnmin fails for probs with both bounds on Jenn_Samp, Meyer, 
     Brown_Den and Osborne_1 (line search failure). Have not tested
     single bounds. Fails Chebyquad (prob 35) if L or U but works B,
     or unconstrained.

   Rvmmin / nvm: 
   o    Saving of final Hessian approx as in Rvmminx
   o    Eventual incorporation of different line searches, possibly in a
        separate routine

   snewtm/snewton:  (These NOT checked as at 20230619)
   o snewtonm does NOT use approximate hessian particularly jacobian(gradient)
     Clean up snewtonmbtest.R ?? check if this done?
   o SNewton.Rmd has examples moved to R scripts -- Need to add approx. Hess.
   o Needed to trap singular Hessian for snewton and snewtm ?? do we?
   o update check of maskmeth i.e., methods that can handle masks, esp for snewtonmb

===================================================================================
NEWS from the optimx package

Some comments apply to deprecated packages optimr on CRAN and optimrx on R-forge.

VERSION 2023-10-21
      o Expand diagnostic for missing or unsuitable method(s)
      o Fix unnecessary ",..." in call to L-BFGS-B in optimr(). The dot-args were 
        removed to avoid an argument collision (VERSION 2023-06-25) using a suggestion
        of Duncan Murdoch, but I forgot to remove this case.
      o Duncan Murdoch proposed and suggested a patch to avoid termination by try(),
        though failure is reported, and to attempt to give better error messages on 
        cause of failure in such cases.
      o Some minor corrections to simple typos and extraneous text.
      o Added checks on valid methods in opm(), trap for bounded methods in optimr().

VERSION 2023-09-01
      o Used Henrik Bengsston's R.rsp method to install static pdf vignettes as some
        CRAN test platforms seem to be missing files needed to build vignettes from 
        Rmarkdown source files. The vignette source Rmarkdown files are in the 
        directory inst/doc/vignette-source.

VERSION 2023-08-23
      o Thanks to Mark Pletscher for uncovering a bug due to a typo in optimr(). Also
        noting that optimr() now fails if input parameters are infeasible. However, 
        users can pre-adjust parameters as follows
           # set initial parameters in vector init, lower and upper bounds in lo and up
           chkpar <- bmchk(init, lower=lo, upper=up)
           if (chkpar$parchanged) init <- chckpar$bvec
      o kktchk now has control with default dowarn=TRUE to allow suppression of overly
        common warning that estimated Hessian at solution is asymmetric. The test for
        asymmetry is very difficult to set up in a way that does not give warnings when
        the Hessian estimate is reasonable (false positives).

VERSION 2023-06-25

      o This is a MAJOR revision and overhaul of the optimx package and its components.
      o offset removed as control. Use reltest.
      o gHgen.R and gHgenb.R are no longer needed by optimx and are kept purely for 
        legacy uses by others.
      o optimr now tries to add function name attribute to returned 'value' object; 
        opm tries to add this attribute to the answer object.
      o 20230730: added pracma::nelder_mead and nloptr::neldermead
      o ctrldefault() now offers collections (vectors) of methods nogrmeth, grmeth, hessmeth
        for those with no gradient requirement, gradient methods and methods that can or
        must have Hessian function supplied
      o opm() does more checks e.g., missing gr, methods inappropriate to problem
      o nvm and Rvmminb: fixed issue that sometimes trystep (distance to bound
        in search direction) came back as small NEGATIVE number) by setting
        any steplength < relstep*.Machine$double.eps to 0. This forces a search
        in gradient (steepest descent) direction and seems to correct runaway 
        backtracking that fails out on too much work.
      o Fixed CITATION file based on R CMD check --as-cran complaints 
        regarding requirement for person() and bibentry() changes.
      o In inst/doc/examples/fhop.R: found we need to ensure both up and lo
        defined to get certain solvers to work e.g., lbfgsb3c. This meant
        adjusting fhop to put Inf or -Inf in as appropriate.
      o checksolver() and checkallsolvers() revised and documented. No 
        check now in ctrldefault() for solver availability. That is, we
        will tolerate a failure and rerun if needed solvers are missing.
      o Added "Intro to optimx" vignette. Specify that gradient function
        should be supplied explicitly if possible.        
      o Add method="MOST" options to try most methods in opm(). This is
        a list returned as item 'mostmeth' from ctrldefault().
      o optimx legacy function fns use numDeriv explicitly for grad, jacobian, 
        and hessian
      o dot arguments added in optimr() to subplex
      o Importance of named arguments over positional ones noted in Intro vignette.
      o Minor fix to test of maximize in optimx.setup()
      o Legacy optimx() demos in inst/doc/legacy-demo/. Note that these have been
        checked to ensure they "run", though results have NOT been verified.
      o Examples in inst/doc/examples/ have all been checked to ensure they run.
        Results have NOT all been verified in detail.
      o A test of 'hessian="approx"' is in vignettes/GenRoseHess.R
      o grchk() is now called in optimr, NOT solvers, except (for the moment)
        in Rvmmin().
      o Example script fhop.R that uses the funconstrain package tests now has
        a check that funconstrain package is installed, and suggests how to
        install from Github since it is not on CRAN.
      o A long-standing issue is that approximated Hessian at return is NOT
        symmetric. i.e., kktchk -- Symmetry check often fails and we symmetrize 
        the computed Hessian. See testing script inst/doc/examples/HessQuality.R
      o demo() has been updated to use newer functions, but is not yet tidy.
      o grpracma.R allows "grpracma" as gradient function and uses pracma approx.
        ctrldefault()  hesspkg and jacpkg controls to use pracma rather than
        numDeriv for hessian and jacobian approximations.
        getFromNamespace("fn","pkgname") to set up functions like jacobian, hessian
      o Checked special controls and adjusted for bobyqa method. Note 
        inst/doc/examples/specctrlhobbs.R
      o Added scounts to optimr, updated proptimr() to use these if available,
        also summary.opm()
      o Using ideas from optim() (thanks to Duncan Murdoch for pointer), dotargs
        suppressed internally in fn, gr, hess to avoid argument clash as in 
        https://stackoverflow.com/questions/69033754/maximum-likelihood-estimation-of-three-parameter-reverse-weibull-model-implement
        Added in optimx() as well as optimr()
      o ncg reimplemented from Rcgmin to use checking done at start of optimr and 
        avoid duplication of effort. ncg() should NOT be used alone.
      o Maximization problems have function value reversed in optimr() NOT opm()
      o Add runex.R for running examples in inst/doc/examples/
      o Stripped down nvm() to use checks already done in optimr()
      o Commented out user input after error in line search in Rtnmin-package.R
      o Fixed minor glitch in bmchk() status indicators
      o Created first try at snewtm() to run only from optimr; snewtonm() left as
        stand-alone function



VERSION 2022-11-25 (was 2022-06-10)

      o kktchk -- Symmetry check often fails and we symmetrize the computed Hessian.
	Note the testing script HessQuality.R (20220610)
      o Rcgmin.Rd has examples changed to properly handle approximations via
        optimr().
      o Added note to opm.Rd that maxfeval is passed through to methods that can
        use it.
      o snewtonmb --> snewtonm while snewtonm -> snewtonmu (unconstrained)
        as bounded verson faster with unbounded cases. Removed snewtonmu.R to 
        00unversioned. Script snewtonmbtest.R has an "all masked" test which works.
        bmchk() modified so it correctly adjusts masks and only sets parchanged
	  when they parameters are adjusted
      o optimr() returns a "status" attribute for bounds which proptimr() displays
        if available
      o control checkgrad is used to invoke function grchk() to check supposedly
        correct (analytic) gradient code.
      o Change bmchk to use an offset for checking equality -- otherwise get poor
        results from "status" indicator on parameters
      o The Argonne tests (More', J. J., Garbow, B. S., & Hillstrom, K. E. (1981).
        Testing unconstrained optimization software. ACM Transactions on Mathematical 
        Software (TOMS)}, {7}(1), 17-41. doi.org/10.1145/355934.355936) in the
        funconstrain package to which J C Nash added Hessians, 
        	https://github.com/jlmelville/funconstrain
        Now have a convenient test script fhop.R
      o Replaced doi in CITATION because R CMD check complained. Possibly JSS has
        changed some specs on its doi values.
      o Added 'weakmeth' to ctrldefault() and these methods do NOT contribute to
        opm() with method="ALL"
      o Many warnings 'pHes not symmetric' in ffn runs. Tried to scale to nHes, but
        unsymmetric hessian gives complex eigenvalues, so NOT a good idea to ignore.
      o added marqLevAlg::mla to optimr. Note extremely different return structure
        from optim() etc.
      o ncg and ncgqs issue with bdstest with masks -- 
        needed to pass bdmsk to solver correctly in optimr
      o returned hessian uses hess() if available, else approx.
      o opm and summary.opm add bounds status.
      o "probname" attribute to ans$value added to output of optimr.
      o Status indicators for parameters (F, L, U) in optimr, proptimr. 
      o proptimr() displays bounds, masks info from bmchk$bchar. optimr updated.
        Also displays solution being C(onstrained) or U(nconstrained)
        opm() NOT yet updated. 
      o kktchk: if pHes not symmetric, symmetrize after warning.
      o created snewtonmb to handle bounds in a Newton-like method.
	This works for a simple function in inst/doc/examples/snewtonmbtest.R 
        where both bounds and masks appear to work for the example.
        However, one correct example is NOT a guarantee of all cases.
        snewtonm should be considered EXPERIMENTAL and DIDACTIC for the present.
        The author/maintainer would like to be informed of uses, successful or
        not, so the code can be improved.
      o nmkb is dropped from set of methods for opm() if start is on a bound.
        There is a test of this in inst/doc/examples/dropnmbk.R
        optimx() does NOT check for this possibility.
      o bmchk reviewed
      o optimx-package.Rd revised
      o Added hessian-uses.R example of Hessian use.
      o Add warning if user tries to use optim::SANN in optimx() or optimr()
      o demo directory removed. Some checks/tests that are to verify code (but not
        needed after code is working) are in inst/doc/examples/
      o added testbmstep.R in inst/doc/examples/ to check bmstep(). Minor addition of
        comments to bmstep.R
      o added starttests to ctrldefault for use in optimx(). There is an example of
        use as trystarttests.R in inst/doc/examples/.
      o opm() now uses unique(method) to avoid duplicates which crash function.
      o Note Rvmminb was missing some bounds resetting steps. 
        Also some cleanup of details for algorithm controls to use ctrldefault()
        e.g., replace dblmax with control$bigval 
      o grchk and hesschk have ffn1 and ggr1 set to avoid dotargs to avoid name
        clash
      o axsearch() now gets controls from ctrldefault(). 
	NOTE: THE CALL SYNTAX HAS CHANGED SLIGHTLY. 
	inst/doc/replace2021/axsearch2018.R has the old version.
      o maxfn -- kkt test fixed hjn hjkb. Call to hjn from optimr 
        uses mcontrol, not control (maximize issue).
      o Some cleanup of minor issues of output messages
      o Rvmmin, Rvmminu, Rvmminb set to same codes as in Rvmmin "package" updated
	December 6, 2021. However in optimr we do NOT use control$maximize.
        Maximization is done via optimr().
        NOTE: there was an algorithmic error in the earlier codes. FIXED.
      o control$checkgrad in Rvmmin defaults to FALSE (was TRUE). In Rcgmin already so.
      o Minor improvements to opm.Rd and optimr.Rd regarding control list.
      o 'convcode' replaced by 'convergence' in opm.Rd
      o lbfgs does NOT report fevals and gevals -- put in optimr documentation
      o opm() adjusted so snewton and snewtonm removed from method list when hess()
        not provided (==NULL)
      o fixed three cases where setting control$maximize=TRUE was failing to set
        final function value correctly
      o removed printout of trace from Rvmminb()
      o minor edits for grammar in snewton.Rd
      o moved some tests to ./inst/doc/examples and tidied the files
      o optimx routines did not check for NULL gradient or hessian routine gr 
        and would fail in some instances. Check for missing routines now included,
        and optimx.setup forces ugr and uhess NULL if necessary.
      o trig1507.R moved from tests to inst/doc/examples/ as the multiple minima 
	generated are unsuitable as tests

VERSION 2020-01-18
      o This version does not call some non-CRAN packages.

VERSION 2019-12-02
      o Replaced all instances of tests on class() with inherits()
      o Replaced calls to lbfgsb3 package with those to lbfgsb3c,
        and removed calls to the lbfgsb3() function
      o Needed to set fval to NA in fnchk when length > 1

VERSION 2018-07-10
      o Added proptimr() compact display of result of optimr
      o Some minor changes to vignettes and to message (msg) output for 
        line ending

VERSION 2018-07-06
      o Changed setup of control vector using ctrldefault to allow for use of
	individual method defaults

VERSION 2018-04-17
      o Comments in tests etc. to explain why methods fail
        Add hessian to simple test
      o Split optim() call in optimx() based on have.bounds. Note glitch in 
	optim() for vector vs scalar infinite bounds.
      o Fix Rvmmin for n=1 infinite loop      

VERSION 2018-03-30
      o Cleanup of output from various solvers and fix some minor omissions
	and errors in the returned information
      o Conditional output of nlm diagnostic info.
      o Changed Rvmmin call. Rvmminb and Rvmminu NOT now called separately,
	as user may inadvertently supply inadmissible parameters and get wrong
	answer.
      o Changed optimr() so nlminb() can use ehess() function
      o Replace convcode in optimr() with convergence to align with optim() result
      o axsearch() example added to bdstest.R in tests.
      o Merged optextras package into optimx
      o rationalized  optimz-package.R, optimx-package.R into one set of code
      o Tests for grchk, kktchk, fnchk included
      o Fixed unconditional printout of maxit
      o savehess saves (approximate!) hessian with nlminb, call with 'hess' not 'hessian'

2018-03-28 in Rtnmin
   o Try ! is.numeric(g) vs is.null(g)
   o trace changed to integer

VERSION 2018-03-18
      o added 
           mcontrol$fnscale <- control$fnscale
        in optim() section of optimr for correction of maximization 
        Thanks to Carlo Lapid cmlapid@up.edu.ph
      o Thanks to playing with examples from Hans Werner Borchers for Nelder-Mead
        style codes, found that optimr() did not transfer control$maxit to the
        optim() functions. Fixed.

2018-03-03
  o update snewton vignette and tests
  o sorted out optimr nlminb failure (used hessian vs hess in optimr call)

VERSION 2017-12-15
      o merging optimr(x) into optimx
      o some documentation of differences between opm() and optimx()
      o Use pkg::fn style calls in optimx.run.R
      o add bib references to vignette
      o checked method="ALL" and control(all.methods=TRUE) are equivalent for opm()
      o set trace=0 in all tests
      o partially simplify tests

VERSION 2017-11-25 optextras
   o    Changed kktchk argument maxfn to maximize to agree with 
        optimr and other packages

2017-11-25 in Rvmmin
   o    Tidy tests
   o    Some cleanup of vignette

VERSION 2017-10-05
      o Ravi Varadhan as maintainer (reverted later)
      o Warning about combining options.

VERSION 2017-09-24
      o comments include 2017-09-18 version
      o brought optimx solvers into optimr and added check that packages 
        are available
      o Fix for lbfgsb3 which was changed at same time to correct and
        simplify some control parameters

2017-7-18 in Rvmmin
   o    Simplify example
   o    Added tests

VERSION 2017-04-04
      o allowed hessian in opm() and optimr() (for nlm)
      o Thanks to Marie Boehnstedt for catching error in creation of
        the composite function and gradient for calling nlm() method.
 
VERSION 2017-02-15
      o added subplex method

2016-11-30 in Rvmmin
   o    Vignette added (incomplete)

2016-09-22  in Rtnmin
   o    account for null gradient 

VERSION 2016-09-16
      o Harry Joe reported troubles with numerical gradient approximations. Found
        some glitches and conflicts to fix. Also documentation is less than ideal.
        Scaling spar to par was INSIDE control$trace

VERSION 2016-08-26
      o Thanks to Henrik Singmann for spotting a bug in optimx.run.R
        for the case where nlminb fails. I had missed resetting some
        elements of the solution object in this case. 

VERSION 2016-07-10
      o hessian=TRUE is NOT properly executed 
      o Test function hobbs15b illustrates different scalings across all methods
      o Appropriate lists of "ALL", "bounded" and "masked" methods are defined 
        in the ctrldefault() function.
      o jonesrun L-BFGS-B fails with numerical gradient approximation.
      o dfoptim hjkb was fixed by Ravi Varadhan

VERSION 2016-07-03
      o Move allmeth, bdmeth, maskmeth into ctrldefault so we have just in
	one place. Note that this will change with optimr/optimrx
      o Masks for hjn(), and incorporate hjn in optimr
      o time (xtimes) removed from optimr, but timing is included in opm()
      o polyalgorithm wrapper polyopt() included and documented with example
      o multistart wrapper multistart() included and documented with example
      o renamed coef.optimx to coef.opm


VERSION optextras 2016-06-30
   o    Streamlined the package to get rid of ufn, ugr, uhess, 
        ugHgenb that are so far unused. 
   o    The optsp environment is only currently used
        with the simple numerical gradient approximations.
   o    kktc renamed to kktchk. scalecheck renamed to scalechk
        for consistency with other naming.

2016-06-28 in Rtnmin
   o    Removed development files from inst, and inst directory

VERSION 2016-06-28
      o Checked that controls do not conflict, e.g., maximize and fnscale. Note that
        maximize is NOT in optim() call.
      o Checked and fixed scaling of bounds in optimr. However, hjkb seems to fail
      o Creates scbdtest.R,  a test function that is simple but bounded and scaled badly
      o Get routines from optextras rather than have duplicates in optimr
      o rename package to optimr for optimR(eplacement)
      o Problems with lbfgs package (unconstrained) resolved
      o test parameter scaling (but this could be improved)
      o Demos still refer to optimx -- must be changed -- 160628 moved 
        off to Roptimtest/current/
      o Decided to postpone adding nloptr to set of solvers

VERSION 2016-05-30
      o Crude setting of uobyqa and newuoa for control rhobeg.
        Especially when parameters all 0, or bounds undefined, program 
	will fail so we need to set this. May also be needed for rhoend.

VERSION 2016-01-07
      o realized control rhobeg needed to be set or bobyqa would fail.

2015-11-30 in Rvmmin
   o    Changed control "stopbadupdate" to TRUE from FALSE.

2015-06-03 in Rvmmin
   o    Had "lower" instead of "upper" in one of the bounds checks.
        Fixed.

VERSION 2015-06-03
      o Added index to bstate in optimx.check.R to avoid unwanted 
        elements in reporting bounds when trace > 0

2015-05-03 in Rvmmin
   o    Check gradient count ig > 2 when doing convergence test.
	Possible failure for test fn = sum(x*x) noted. However, 
	other fns fail with this change. This is an issue in the
	program logic rather than an algorithmic failure. 

VERSION 2015-04-23
      o removed condition that made solution NA if gradient or function
        limits exceeded. Now returns parameters and function at the point
	where limit reached.

VERSION 2015-01-21
      o Added lbfgsb3, the 2011 version of this code.

2015-01-07 in Rvmmin
   o    comments in codes Rvmminb.R and Rvmminu.R aligned so that the
        differences between bounds constrained version and unconstrained
        more easily visualized
   o    control keepinputpar explained. It is used when user does NOT want
        out-of-bounds parameters on input moved to nearest bound. Program
        will stop in this case.
   o    Minor changes to maxit and maxfeval defaults so that they are 
        increased with increasing numbers of parameters.


2014-12-21 in Rvmmin 
   o    As per CRAN policies, edited DESCRIPTION and NAMESPACE files. 
	Because there is no change to code, the date of the package 
	advanced by only 1 day.

2014-7-17 in Rcgmin
   o    Changed default of control$checkgrad to FALSE. Changed error message
	that mentions Rvmmin to Rcgmin.

2013-11-11  in Rtnmin
   o    First version of Rtn.

VERSION 2013-10-27 optextras
   o    Inserted missing index in bmchk (bdmsk)

2013-10-27 in Rvmmin 
   o    Added maxfevals to control. (It had been there some time.)


2013-09-12 in Rtnmin
   o    Modlnp fix from SGN implemented so jones function works.

VERSION 2013-9-11 optextras
   o    Fixed some elements of manual. Alias optextras-package
        to optextras.

2013-09-03 in Rtnmin
   o    Some fixes, particularly to dot-arguments missing
	from calls to user function.
   o    Changed returned user fgfun to f with g as attribute.

VERSION 2013-8-20 optextras
   o    optsp scratchpad environment added. Note that it
        is aliased in grfwd to avoid a warning in R CMD
        check.
   o    Can change this in numerical derivative routines
        in case of parallel running.

2013-8-19 in Rvmmin
   o    Numerical gradient from optextras, remove bmchk from package,
	put bmchk call in Rvmmin, not Rvmminb

VERSION 2013-08-08
      o Attempt to sort out Rvmmin and Rcgmin no gradient failures. 
        Resolved by REQUIRING a gradient routine. In optimr() can provide
        an approximation via quoted gradient function name.
      o Fix for nlm calls which sometimes failed due to inappropriate
        control settings and failure to clear (NULL) temporary settings.

VERSION 2013-08-07
      o Moved optimizer packages to Imports and removed require() calls to
	satisfy R 3.1.2.
      o This version is artificial! It is the version of 2013-08-06 with an 
	updated CITATION and minor fixes to satisfy NAMESPACE requirements.
	A new optimx is in process with quite sweeping changes that may or
	may not result in changed behaviour and usage. (JN 2014-09-04).
      o Failure to have the following packages installed may give rise to 
	diagnostic messages that can be confusing: Rvmmin, Rcgmin, ucminf,
	minqa, BB, and dfoptim.

VERSION 2013-08-02
      o Change startup to avoid warning messages for missing methods.
        Add function to check which methods available.
      o UOBYQA is present in optimx.run.R, but is not included in the list of
        "all.methods" in optimx.setup.R. It should be possible to run UOBYQA
        by including "uobyqa" in the list of methods in a direct call to optimx.run.R.

2013-7-20 in Rvmmin
   o    Minor improvements to timing by use of crossprod() in 
        computing y and D2

VERSION 2013-07-09
      o patch for missing "details" behaviour
      o point to example of use of "follow.on" control 

VERSION 2013-07-03 
      o added control usenumDeriv

VERSION 2013-06-04 
      o improved coef() function to extract parameters

2013-06-03 in Rtnmin
   o    Properly packaged version of Rtnmin. 

VERSION 2013-05-03
      o removed optansout.R, get.result.R, get.best.R (no longer needed)

VERSION 2013-04-30
      o added optimx<-.coef

VERSION 2013-04-29
      o removed help pages documenting obsolete features

VERSION 2013-04-06
      o nmkb and hjkb had no ... arguments in call. Fixed.
      o L-BFGS-B (and possibly other optim() methods) return no gevals (count[[2]])
        on failure. Forced to NA in case of try-error.

VERSION 2013-04-05
      o Fixed maximize of function without gradient function. Also glitch
        (not tested or encountered) with user supplied Hessian function when
        maximizing.
      o Separate .Rd files for coef() and summary()
      o Use have.bounds to select bounds and unconstrained methods where there
        is a choice to avoid duplication of tests. This has been done for 
	hjk, nmk, Rcgmin, Rvmmin
      o Revision of output description in optimx.Rd; some simplifications.
      o Parameter names preserved from starting vector, or "p1", "p2", etc. used.
        There is a test for these in the ox demo().

VERSION 2013-04-03
      o in summary.optimx no re-ordering is now specified using order = NULL
      o Fixup of examples -- simplified, corrected call to details

VERSION 2013-04-02
      o in summary.optimx the order= argument may now be an unevaluated 
        expression (as before), a string representation of that expression or 
        a character vector of names, e.g. these all work:
           order = value # unevaluated expression
           order = "value" # character string containing column name
	   order = list(fevals, gevals) # unevaluated list
           order = c("fevals", "gevals") # character vector of column names
           order = list(round(value,3), fevals) # unevaluated list
           order = "list(round(value,3),fevals)" # expression as string

VERSION 2013-03-29
      o Note that asking for a method that is NOT in the results will return
        a row with rowname (the method name) as "NA". Note that this will NOT
        test TRUE with is.na() on the rowname. Use instead a test like
               (row.names(ans8missmeth["Rvmmin",])[[1]] == "NA")
      o kkt1 and kkt2 in results now returned TRUE/FALSE as in 2012 versions
      o Fixed error in kkt2 test (using old name evnhatend rather than hev)
      o Results are no longer sorted by optimx, but can be using the 'summary()'
        function. This can be called in a variety of ways. 
      o The 'details' are always kept in a matrix form, even when only one
        method has been used.
      o 'convcode' is used in place of 'convergence' in result structure
      o There is a directory inst/interactive-test that contains ox.R test
        script that pauses so user can see interim results. 
      o As a result of several issues, the result structure is changed from
        that of optimx 2012 versions. It is now a data frame with a `details'
        attribute. Also an 'npar' attribute to give the number of parameters,
        and a 'maximize' attribute that is TRUE when the function is to be
        maximized.


VERSION 2013-03-28
      o print.optimx dropped. Not needed as print defaults to print.data.frame.
      o added summary.optimx which has order= and par.select= arguments
      o order= is an expression or list of expressions using the column names 
        of the "optimx" object as variables (plus the variable rownames).
      o par.select = FALSE now has same meaning as par.select = 0 (rather than 
        being an error)
      o [.optimx now subsets details attribute too
      o method column from details attribute dropped and moved to row names

VERSION 2013-03-27
      o created unit test framework with 2 unit tests in 
          inst/unitTests/runit.all.R.  
        To run: demo("unitTests")

VERSION 2013-03-25
      o [.optimx and as.data.frame.optimx added
      o coef.optimx fixed to reflect new ans.ret

VERSION 2013-03-22
      o ans.ret structure changed

VERSION 2013-03-21
      o maximize works with print(x, best.only=TRUE)
      o KKT checks can be switched off. Decided to NOT separate the code from
        optimx.run  UNTESTED
      o optimx.setup, optimx.run, optimx.check are NOT exported in NAMESPACE,
        but a knowledgeable user could use these codes directly, either from
        the source code or by modifying the package locally.

VERSION 2013-03-18
      o removed method= and columns= arguments from print.optimx and added
        best.only
      o removed print.optimx help page and moved it to the optimx help page
      o the row names of the optimx output now show the method and the method
        column has been removed

VERSION 2013-03-18
      o Removed all user accessible functions except optimx and added
        print.optimx.

VERSION 2013-03-16
      o  get.result now emits a warning if method requested is not in
         optimx solution
      o  attempt 2013-3-17 to include summary() in methods to replace
         trimpars()

VERSION 2012-07-18 optextras
   o    Extracted helper routines from optplus to 
        create optextras package to put on CRAN.
        Formerly in optplus or optfntools on R-forge,
        but difficulties with dependencies indicated
        that the support routines need to be on CRAN.
        Renamed to avoid name confusion.

VERSION 2012-07-09 optextras
   o    Suppress warnings in ugHgenb -- too much
        unhelpful output
   o    Fixed optansout to print parameters

VERSION 2012-06-07 optextras
   o    Merged optfntools into optplus to ensure 
        R-forge build.

VERSION 2012-05-26 optextras
   o    Added check for "on a bound" in bmchk to aid
	in use of tanh() transformation as in nmkb().
   o    scalechk takes account of masks (masked parameters not used for 
        computing scale)
   o    Changes before 2012-5-26 in the optextras package not recorded 

2012-05-21 in Rvmmin
   o	Separation of constrained and unconstrained versions.

2016-6-30
   o    Streamlined the package to get rid of ufn, ugr, uhess, 
        ugHgenb that are so far unused. 
   o    The optsp environment is only currently used
        with the simple numerical gradient approximations.
   o    kktc renamed to kktchk. scalecheck renamed to scalechk
        for consistency with other naming.

2013-9-11
   o    Fixed some elements of manual. Alias optextras-package
        to optextras.

2013-8-20
   o    optsp scratchpad environment added. Note that it
        is aliased in grfwd to avoid a warning in R CMD
        check.
   o    Can change this in numerical derivative routines
        in case of parallel running.

2012-7-18
   o    Extracted helper routines from optplus to 
        create optextras package to put on CRAN.
        Formerly in optplus or optfntools on R-forge,
        but difficulties with dependencies indicated
        that the support routines need to be on CRAN.
        Renamed to avoid name confusion.

2012-7-09
   o    Suppress warnings in ugHgenb -- too much
        unhelpful output
   o    Fixed optansout to print parameters

2012-7-9 in Rvmmin
   o    Removed optansout to avoid conflicts with optplus/optimx
   o    Changes before 2012-4-30 in Rvmmin not recorded 

2012-6-07
   o    Merged optfntools into optplus to ensure 
        R-forge build.

2012-5-26
   o    Added check for "on a bound" in bmchk to aid
	in use of tanh() transformation as in nmkb().

2012-5-21 in Rcgmin
   o	Separation of constrained and unconstrained versions.
   o    Changes in Rcgmin before 2012-4-30 not recorded 



TODOS in Rcgmin or ncg or Rvmmin or nvm
   o    Eventually include different line searches, possibly in a
        separate routine

2013-9-24  in Rcgmin
   o    Warning that parameter moved to bound suppressed when start
        is on the bound.

2013-8-20 in Rcgmin
   o    Allow quoted gradient names to bring in different gradient 
        routines. Also to allow gradient approximation step to be
	adjusted.
   o    Dependency on optextras rather than numDeriv

2013-2-20 in Rcgmin
   o    Removed fail, setstep (not used), ceps from Rcgminu (not used)

2012-8-08 in Rcgmin
   o    Minor edits and addition of references.

2012-6-09 in Rcgmin
   o    User-tolerance for gradient size.
===================================================================================