Sun 5 Oct 2025 : removed excluded_source_files from continuation.gpr.

Mon 25 Dec 2023 : added verbose level to *_blackbox_refiners.a* and
  adjusted calls in *_blackbox_continuations.adb.
  After adding verbose level in {standard,dobldobl,quaddobl}_root_refiners,
  adjusted calls in *_blackbox_refiners.adb.

Thu 10 Jun 2021 : added the exec_dir attribute to continuation.gpr.

Wed 9 Jun 2021 : defined continuation.gpr to build the test procedures.

Tue 23 Feb 2021 : updated ts_poco.adb for the renaming of the package
  drivers_for_poly_continuation into main_poly_continuation.

Sun 13 Dec 2020 : to {standard,dobldobl,quaddobl}_blackbox_continuations,
  added procedures to provide the gamma for Laurent systems.

Sun 6 Dec 2020 : moved mainpoco into main_homotopy_continuation.

Sun 22 Nov 2020 : moved bablpoco into standard_blackbox_continuations,
  bablpoco2, bablpoco4 into {dobldobl,quaddobl}_blackbox_continuations.

Sun 8 Nov 2020 : renamed drivers_for_poly_continuation into main_poly_
  continuation, adjusted mainpoco.adb.

Sun 1 Nov 2020 : added test on Is_Null(sinsols) in {dobldobl,quaddobl}_
  blackbox_refiners.adb to avoid a crash with phc -b2 -t and phc -b4 -t
  in case the system has no singular solutions.

Tue 6 Aug 2019 : modified drivers_for_poly_continuation.adb so prclvl = 0 also
  stands for double precision, as a result of changing default from 0 to 1.

Thu 20 Jun 2019 : added verbose level to the main driver procedures
  in drivers_for_poly_continuation.

Wed 19 Jun 2019 : added verbose level to *_blackbox_continuations,
  updated bablpoco.adb, bablpoco2.adb, and bablpoco4.adb.

Fri 14 Jun 2019 : added verbose level to the procedures in mainpoco.adb.

Thu 13 Jun 2019 : added the verbose level to mainpoco, bablpoco, bablpoco2,
  and bablpoco4.

Wed 2 Jan 2019 : fixed error in ts_bbpoco.adb.

Tue 18 Dec 2018 : added writing of time stamps to mainpoco.adb.

Fri 31 Aug 2018 : filtered the solutions on reaching the target in
  {standard,dobldobl,quaddobl}_blackbox_refiners.adb.

Fri 6 Apr 2018 : added deflate options to standard_blackbox_refiners
   and standard_blackbox_continuations, which changed bablpoco.adb.

Thu 5 Apr 2018 : added Is_Null() tests in standard_blackbox_refiners.adb,
  in particular for the reporting case when there are no singular solutions.

Wed 4 Apr 2018 : updated standard_blackbox_refiners.adb, removing the writing
  of the refined singular solutions, as it is no longer needed.

Tue 3 Apr 2018 : reverted changes back in standard_blackbox_refiners,
  applying reporting_refiner in the reporting blackbox multitasked refiner.

Sat 31 Mar 2018 : in standard_blackbox_refiners, applied multitasking
  root refinement only to the regular solutions.

Fri 30 Mar 2018 : in {standard,dobldobl,quaddobl}_blackbox_refiners.adb,
  replace singular filter by a splitter.
  Adjusted standard_blackbox_continuations.adb.

Thu 29 Mar 2018 : new packages {standard,dobldobl,quaddobl}_blackbox_refiners,
  with code from {standard,dobldobl,quaddobl}_blackbox_continuations.adb.
  Added more to {standard,dobldobl,quaddobl}_blackbox_refiners.

Wed 28 Mar 2018 : to {standard,dobldobl,quaddobl}_blackbox_continuations.adb,
  added the writing of elapsed wall clock times for multitasked continuation.

Sat 18 Mar 2018 : fixed ts_poco.adb for changed driver procedure in
  drivers_for_poly_continuation.

Sat 19 Aug 2017 : added parameters to drivers_for_poly_continuation and
  updated mainpoco.adb so the end solutions are written to separate file
  in case the precision is higher than double.  Added the root refiners
  to the calling of phc -p in mainpoco.adb so the solutions are written
  properly to the output file.

Wed 28 Jun 2017 : check added in standard_incfix_continuation.adb to see
  if all components of a vector are valid (that is: non NaN).

Sun 8 May 2016 : bug fix in quaddobl_blackbox_continuations.adb with
  the root refiner, to filter the diverging end points at the paths.

Thu 24 Mar 2016 : adjusted {dobldobl,quaddobl}_blackbox_continuations.adb
  so deflation is applied in the root refiners.

Sun 20 Mar 2016 : added root refiners for overdetermined path tracking
  in mainpoco.adb.

Sat 19 Mar 2016 : placed stub for parameter/sweep homotopies of Laurent
  polynomial systems somewhat deeper in mainpoco.adb.

Tue 15 Mar 2016 : updated bodies of drivers_for_poly_continuation with
  additions to the specifications as well.  For Laurent systems, called
  the correct driver, so path tracking on overdetermined systems works!

Mon 14 Mar 2016 : made interface in drivers_for_poly_continuation for
  Laurent systems the same as for regular systems, changed mainpoco.adb.

Sat 12 Mar 2016 : improvements in drivers_for_poly_continuation for the
  tracking of paths defined by overdetermined Laurent homotopies.

Wed 9 Mar 2016 : updated mainpoco.adb to accept overdetermined Laurent
  polynomial systems as input of phc -p.

Tue 8 Mar 2016 : preset precision level passed through drivers to the
  multitasked path trackers, in mainpoco.adb.  Added extra tests in
  drivers_for_poly_continuation.adb, mainpoco.adb calls multitasked
  secant homotopy for Laurent homotopies.

Sat 5 Mar 2016 : adjustments in mainpoco.adb for the multitasking path
  trackers to apply the Gauss-Newton correctors.

Thu 3 Mar 2016 : updates in drivers_for_poly_continuation.adb and
  {standard,quaddobl}_incfix_continuation.adb for overdetermined trackers.

Wed 2 Mar 2016 : added extra parameter in drivers_for_poly_continuation,
  for the overdetermined homotopies.

Mon 29 Feb 2016 : added extra parameter to dobldobl_incfix_continuation,
  which changed drivers_for_poly_continuation.adb and
  dobldobl_blackbox_continuations.adb.

Sun 28 Feb 2016 : updated mainpoco.adb and drivers_for_poly_continuation.adb
  so that phc -p can handle overdetermined homotopies in double precision.

Sat 27 Feb 2016 : extra argument in standard_incfix_continuation,
  which changed calls in standard_blackbox_continuation.adb and
  drivers_for_poly_continuation.adb.

Sun 21 Feb 2016 : added more to the stub procedure in mainpoco.adb.

Sat 20 Feb 2016 : added an extra procedure to prepare for overdetermined
  path tracking in mainpoco.adb

Wed 13 Jan 2016 : to drivers_for_poly_continuation added procedures to
  redefine the homotopies, resetting the relaxation power to one.

Mon 11 Jan 2016 : modified drivers_for_poly_continuation.adb to write end
  solutions after the application of polyhedral end games in double double
  and quad double precision.

Fri 8 Jan 2016 : populated the stubs in dobldobl_incfix_continuation.adb
  and quaddobl_incfix_continuation.adb with code.

Thu 7 Jan 2016 : changed standard_incfix_continuation.adb for the addition
  of the estimated winding number as extra argument.  Did the same for
  dobldobl_incfix_continuation.adb and quaddobl_incfix_continuation.adb.
  
Wed 6 Jan 2016 : extended {dobldobl,quaddobl}_incfix_continuation with
  procedures to compute the directions of the paths.
  Updated drivers_for_poly_continuation.adb with those procedures.
  Added extra parameter for the estimated winding numbers in the
  {standard,dobldobl,quaddobl}_incfix_continuation packages called 
  in drivers_for_poly_continuation.adb.

Mon 4 Jan 2016 : in continuation_parameters.adb, fixed setting of the
  predictor type in case of a polyhedral end game.
  Added print statements in standard_incfix_continuation.adb to mark 
  the beginning of the polyhedral end game for a particular path.

Fri 11 Dec 2015 : fixed ts_poco.adb, for preset precision level in driver.

Wed 18 Nov 2015 : updated mainpoco.adb so prompting for precision calls the
  function in communications_with_user.  Added extra argument to take into
  account the precision level set at the command line of phc -p, e.g. as
  in phc -p2 or phc -p4.  In drivers_for_poly_continuation, restored the
  default tuning of the continuation parameters (regardless of precision)
  and took the preset value of the precision level into account.

Tue 27 Oct 2015 : for consistency, renamed black_box_poly_continuations into
  standard_blackbox_continuation, which changed bablpoco and ts_bbpoco.
  New bablpoco2 and bablpoco4 for blackbox continuation in double double
  and quad double precision.  Extended mainpoco.adb so that phc -p can do
  a parameter or a sweep homotopy with double doubles and quad doubles.

Sat 24 Oct 2015 : changed mainpoco.adb for renaming of systems_with_parameters
  into standard_parameter_systems.

Sat 10 Oct 2015 : called the multitasking root refiners in the packages
  {dobldobl,quaddobl}_blackbox_continuations.adb.

Thu 8 Oct 2015 : extended pack_continuation_parameters with get_value and
  set_value to get and set individual values, tested by ts_pactun.adb.

Sun 4 Oct 2015 : updated mainpoco.adb so that phc -p writes the version number
  and the seed used in the random number generators.

Fri 2 Oct 2015 : in {dobldobl,quaddobl}_blackbox_continuations.adb changed
  the values for the tolerances for the black box root refiners.

Thu 1 Oct 2015 : {dobldobl,quaddobl}_blackbox_continuations.adb apply the
  silent root refiners that filter out the path failures.
  Changed {dobldobl,quaddobl}_blackbox_continuations.adb because of
  changed type of the tolerance parameters in the root refiners.

Wed 30 Sep 2015 : settings in {dobldobl,quaddobl}_blackbox_continuations.adb
  were too severe, lessened to the default values for double precision.

Tue 29 Sep 2015 : tuned continuation parameters for double double and quad
  double in {dobldobl,quaddobl}_blackbox_continuations.adb.

Fri 25 Sep 2015 : from {dobldobl,quaddobl}_blackbox_continuations.adb
  remove the statement to write a superfluous solutions banner;
  added calls to the root refiners, still without multitasking.

Sun 6 Sep 2015 : in {dobldobl,quaddobl}_blackbox_continuations.adb, the
  multitasked Laurent path trackers are enabled.

Sat 5 Sep 2015 : new packages {dobldobl,quaddobl}_blackbox_continuations,
  with new test program ts_bbpoco.

Thu 19 Feb 2015 : fixed bug in printing relative residuals in process_io.adb.

Tue 26 Aug 2014 : extended {dobl,quad}dobl_continuation_data_io with
  procedures to write path tracking statistics.

Fri 15 Aug 2014 : added pWrite operations for quad double in process_io.

Fri 11 Jul 2014 : added option to the drivers_for_continuation_parameters
  to add the precision level, in the file drivers_for_poly_continuation.

Tue 24 Jun 2014 : drivers_for_poly_continuation.adb use coefficient homotopy.

Mon 23 Jun 2014 : black_box_poly_continuations.adb use coefficient homotopy.

Fri 13 Jun 2014 : lowered default number of corrector iterations to 3,
  instead of 4 in continuation_parameters.ads.

Sun 23 Mar 2014 : added _with_Stop routines to the packages
  {dobldobl,quaddobl}_incfix_continuation, as needed in samplers for
  the homotopy membership tests.

Wed 26 Jun 2013 : updated mainpoco.adb for natural32 and integer32.

Tue 25 Jun 2013 : updated black_box_poly_continuations for integer32.

Wed 19 Jun 2013 : changed mainpoco for natural32.

Wed 12 Jun 2013 : about half of all files changes because of natural32
  and integer32 types.

Thu 16 May 2013 : changes in continuation_parameters and its _io.adb.

Wed 15 May 2013 : set the quadratic predictor as default in the
  continuation_parameters, changed continuation_parameters_io.adb.
  In black_box_poly_continuations.adb, changed Tune(2) back to Tune(0).

Fri 10 May 2013 : added multitasking root refiners to the
  black_box_poly_continuations.adb.

Sat 16 Feb 2013 : changed mainpoco.adb for multitasking dd & qd tracking
  and exported Ask_Symbol in drivers_for_poly_continuation.

Fri 8 Feb 2013 : updates in mainpoco.adb and drivers_for_poly_continuation
  for multiprecision path tracking.  Updates in ts_poco.adb.

Thu 7 Feb 2013 : new test program "ts_autotune.adb" to get a feeling of
  the automatic tuning of the continuation parameters.

Tue 5 Feb 2013 : tol_zero in {dobldobl,quaddobl}_incfix_continuation.adb
  must depend on the settings in the continuation parameters.

Mon 21 Jan 2013 : moved code to the new directories ../End_Games, ../Sweep,
  and ../Trackers for clarity.

Mon 8 Aug 2011 : added to standard_continuation_data a procedure to
  compute and write path tracking statistics.

Wed 22 Jun 2011 : added function to process_io.

Thu 9 Jun 2011 : changed black_box_poly_continuations.adb, mainpoco.adb, 
  and parameter_homotopy_continuation.adb
  for changed specifications in standard_root_refiners.

Thu 2 Jun 2011 : added output to {dobldobl,quaddobl}_incfix_continuation.adb
  and updated drivers_for_poly_continuation.
  Added "working_precision" as a variable in continuation_parameters.

Wed 1 Jun 2011 : changed process_io.adb, multprec_incfix_continuation.adb
  multprec_predictors.adb, {dobldobl,quaddobl}_vlprs_algorithm.adb
  for warnings; further work on drivers_for_poly_continuation.
  Changed driver_for_winding_numbers.adb for warnings.

Tue 31 May 2011 : updated drivers_for_poly_continuation,
  tested by ts_poco.adb.

Mon 30 May 2011 : for warnings, changed {dobldobl,quaddobl}_path_trackers.adb,
  standard_quad_turn_points.a* and directions_of_dobldobl_paths.adb.
  New packages {dobldobl,quaddobl}_incfix_continuation change
  drivers_for_poly_continuation.adb, tested by ts_poco.
  Fixed bug in drivers_to_track_{dobldobl,quaddobl}_paths.adb.
  Added "Write_Solution" to multprec_incfix_continuation.adb.
  Formatting code in standard_continuation_data_io.adb.

Sun 29 May 2011 : for warnings, changed drivers_for_poly_continuation.adb,
  only 1 warning left in increment_and_fix_continuation.adb.
  Modified standard_path_trackers.adb to remove one warning.
  Changed valipoco.adb for warnings.

Mon 23 May 2011 : added driver to drivers_for_poly_continuation.
  Upgraded drivers_to_track_{dobldobl,quaddobl}_paths for total degree
  start systems, changed drivers_to_track_paths.adb as well.

Fri 20 May 2011 : fixed type error in drivers_to_track_paths.ads.
  New packages: drivers_to_track_{dobldobl,quaddobl}_paths,
  {dobldobl,quaddobl}_continuation_data_io, tested by ts_track.
  Changed ts_path.adb because of so many warnings.

Sun 14 Nov 2010 : updates in drivers_for_poly_continuation, ts_path,
  and continuation_parameters for path tracking with double doubles,
  touched quaddobl_predictors.adb to trace bug in quad_double_numbers.

Sat 9 Oct 2010 : added Tune to ts_correct.adb to adjust tolerance for
  corrector depending on the working precision.  Changed the creators in
  {standard,dobldobl,quaddobl}_continuation_data.adb to update (err,rco,res).

Fri 8 Oct 2010 : updates in ts_path.adb for path tracking in double doubles,
  made modifications in process_io for quad doubles.  Changed
  directors_of_quaddobl_paths.adb.

Thu 7 Oct 2010 : added stubs to dobldobl_correctors, modified process_io
  for writing info during path tracking with double doubles.
  Altered directions_of_dobldobl_paths.adb, ts_path.adb compiles!

Wed 6 Oct 2010 : extended ts_preco.adb with tests on dobldobl_predictors,
  and on the new quaddobl_predictors, created dobldobl_dispatch_predictors
  and quaddobl_dispatch_predictors.  Renamed directions_of_solution_paths
  into directions_of_standard_paths, to replace the standard by dobldobl
  and quaddobl, created dobldobl_path_trackers, tested by ts_path.adb.

Wed 22 Sep 2010 : created {dobldobl,quaddobl}_vlprs_tables, tested for
  compilation by ts_vlprs.adb, upaded standard_winding_numbers.adb,
  created {dobldobl,quaddobl}_vlprs_algorithm
  and {dobldobl,quaddobl}_winding_numbers.

Tue 21 Sep 2010 : fixed compiler errors in dobldobl_predictors.adb.

Mon 20 Sep 2010 : created dobldobl_predictors, tested by ts_preco.adb.

Sun 19 Sep 2010 : created quaddobl_continuation_data, and updated the
  dobldobl_continuation_data, updates in dobldobl_correctors.adb and
  created the new quaddobl_correctors, tested by ts_correct.adb.

Sat 18 Sep 2010 : updated ts_correct.adb.

Wed 15 Sep 2010 : added code to dobldobl_correctors.adb.

Tue 14 Sep 2010 : types {pred,corr}_pars moved from standard_continuation_data
  into continuation_parameters.  This move affected *_continuation_data.ads
  and multprec_continuation_data.adb, {standard,dobldobl}_correctors.ads,
  ts_correct.adb, ts_preco.adb, standard_path_trackers.*,
  multprec_path_trackers.adb, standard_dispatch_predictors.a*,
  drivers_to_track_paths.adb, and also updated ts_exp.adb, ts_vlprs.adb.

Mon 13 Sep 2010 : created dobldobl_continuation_data, for use in the new
  dobldobl_correctors, tested by ts_correct.adb.

Tue 10 Aug 2010 : new package standard_monomial_correctors, tested by
  ts_correct.adb.

Wed 2 Jun 2010 : because phc -p offers no support for Laurent systems,
  changed drivers_for_poly_continuation.a*, black_box_poly_continuations.adb,
  and mainpoco.adb.

Fri 6 Nov 2009 : changed drivers_for_path_directions.adb,
  solutions_with_parameters.adb, multprec_correctors.a*,
  black_box_poly_continuations.adb and standard_path_trackers.a*,
  standard_predictors.a*, continuation_parameters_io.adb,
  for GNAT GPL 2009 warnings.

Fri 9 Oct 2009 : extended black_box_poly_continuations with continuation for
  Laurent polynomial systems.

Wed 30 Sep 2009 : commented out exception handlers in
  black_box_poly_continuations.adb.

Tue 29 Sep 2009 : changes in black_box_polynomial_continuations.adb for some
  warnings and to fix bug in deflation.

Thu 10 Sep 2009 : exception handlers to black_box_polynomial_continuations.adb
  to check crash with phc -0 -b cyclic7 /tmp/output

Fri 3 Jul 2009 : updated parameter_homotopy_continuation.adb for a better
  interface to the sweep homotopies.
  
Wed 1 Jul 2009 : updated standard_quad_turn_points.a* and its _io.*,
  used by standard_quad_sweepers.adb, tested by ts_realcont.adb.

Fri 19 Jun 2009 : parameter_homotopy_continuation.adb now prints out the
  solutions at the end of a sweep.

Wed 17 Jun 2009 : updated black_box_poly_continuations.a* with routines
  for multitasking.

Fri 15 May 2009 : updated parameter_homotopy_continuation.adb because of the
  new package standard_quad_sweepers.  For tasking mainpoco.a* was changed.

Mon 11 May 2009 : splitted off code from standard_quad_turn_points to
  the new package standard_quad_sweepers.

Tue 5 May 2009 : updated standard_quad_turn_points.a* with a bisection
  method for the noon4A case when the determinant flips sign.

Mon 4 May 2009 : updated standard_quad_parameters.a*,
  standard_quad_turn_points.adb, parameters_homotopy_continuation.adb,
  and ts_realcont.adb.

Fri 1 May 2009 : updated standard_quad_turn_points.a* and its *_io.a*
  for the backup with quadratic interpolation

Tue 14 Apr 2009 : moved "Write_Sweep_Summary" from the "Parameter_Homotopy_
  Continuation" package to "Standard_Quad_Turn_Points_io" for use in the
  ts_realcont.adb.  Updated standard_quad_parameters.ads and also
  standard_quad_turn_points.adb to backup at a critical point.

Mon 13 Apr 2009 : moved "Read_Positive_Float" from "Continuation_Parameters_io"
  into "Numbers_io" of the "Math_Lib/Numbers" directory.
  New package "standard_quad_parameters" manages values of the numerical
  parameters for the sweep, used in standard_quad_turn_points, with tuning
  done by ts_realcont.adb.  Also parameter_homotopy_continuation.adb changed.

Tue 10 Mar 2009 : updated standard_quad_turn_points.a* and ts_realcont.adb
  so user no longer has to provide start solutions.

Fri 15 Aug 2008 : blackbox solver must write homotopy parameters (gamma)
  to file, changed black_box_poly_continuations.adb to do this.

Thu 14 Aug 2008 : put "toggle" commments in black_box_poly_continuations
  to toggle off deflation, for bugs with trinomials...

Tue 1 Jul 2008 : in black_box_poly_continuations.adb fixed crash when 
  there are no general start solutions.

Sun 13 Apr 2008 : in black_box_poly_continuations.adb, added "flush"
  statements before the root refinement phase.
 
Thu 14 Feb 2008 : added counters to the sweep routines in
  parameter_homotopy_continuation.

Tue 12 Feb 2008 : fixed bug in sweep summary 
  of parameter_homotopy_continuation.adb.

Fri 8 Feb 2008 : in parameter_homotopy_continuation, a new procedure
  writes a sweep summary a the end.

Mon 21 Jan 2008 : commented out those exception handlers that were not
  deemed effective in standard_predictors.adb, standard_correctors.adb,
  standard_path_trackers.adb, and increment_and_fix_continuation.adb.

Sun 20 Jan 2008 : fixed some indendation formatting in 
  increment_and_fix_continuation.adb, added exception handlers in
  standard_predictors.adb, standard_correctors.adb, and
  standard_path_trackers.adb, for when -gnatp is not switched on.

Thu 10 Jan 2008 : removed from parameter_homotopy_continuation.adb
  the root refining because it turned out too fragile.

Wed 9 Jan 2008 : changed standard_quad_turn_points.a* and
  parameter_homotopy_continuation.a* to allow to avoid extra output.

Mon 7 Jan 2008 : changed black_box_poly_continuations.adb in search for
  the cause of a crash.  Modified mainpoco.adb because of warnings,
  added root refiners to sweep in parameter_homotopy_continuation.a*.

Thu 3 Jan 2008 : because of warnings in GNAT GPL 2006, changed
  standard_continuation_data.adb and drivers_for_poly_continuation.adb.

Wed 2 Jan 2008 : changed standard_correctors.ads, and
  systems_with_parameters.adb, black_box_poly_continuations.a*
  because of warnings in GNAT GPL 2006.

Sat 29 Dec 2007 : changed standard_quad_turn_points and
  parameter_homotopy_continuation, both .ads and .adb files.

Fri 28 Dec 2007 : to make the sweep also accessible through ts_parcon,
  changed parameter_homotopy_contination, ts_parcon.adb and mainpoco.adb.

Thu 27 Dec 2007 : to prepare the phc -p for the real sweep, modified
  systems_with_parameters, solutions_with_parameters,
  parameter_homotopy_continuation, and mainpoco.adb.

Sat 8 Dec 2007 : changed multprec_path_trackers.adb because of warnings.
  Revamped black_box_poly_continuations to deal with start solutions
  with zero components.

Wed 26 Sep 2007 : ts_realcont.adb reads solutions from input file.

Tue 25 Sep 2007 : minimizing parabolic determinant interpolation added
  to standard_quad_turn_points.adb.

Mon 24 Sep 2007 : improved eigenvalue computation in 
  standard_quad_turn_points.

Mon 10 Sep 2007 : added switch to standard_quad_turn_points.a* to turn
  off the eigenvalue computation, also ts_realcont.adb changed.

Mon 3 Sep 2007 : added monitoring of eigenvalues along path, changed
  standard_quad_turn_points.a*. 

Sun 2 Sep 2007 : added monitoring of determinant along path, changed
  ts_realcont.adb, standard_quad_turn_points + i/o.

Tue 28 Aug 2007 : added some options to ts_realcont.adb, and changed
  standard_quad_turn_points.adb.

Sun 1 Jul 2007 : more changes to standard_quad_turn_points.a* and
  ts_realcont.adb, does not work for quadruple points...

Fri 29 Jun 2007 : prepared standard_quad_turn_points.a* to pass through
  quadratic turning points, also modified ts_realcont.adb.

Thu 28 Jun 2007 : for a more useful ts_realcont.adb, changed 
  standard_quad_turn_points* and its i/o.

Fri 22 Jun 2007 : to have dimension of system written to file, made changes 
  in mainpoco.adb and drivers_for_poly_continuation.adb.
  In standard_path_trackers.adb, success is now an input parameter for
  the single predictors, so that now the Hermite predictor works.

Tue 22 May 2007 : updated black_box_poly_continuations.adb to deflate
  by default in root refinement.

Fri 4 May 2007 : x^2 = 0 and y^2 = 0 made scaling crash in the file
  black_box_poly_continuations.adb, discovered via phc_solve.

Sat 7 Apr 2007 : prepared standard_quad_turn_points for adaptive step size.

Wed 4 Apr 2007 : added "Shoot_Turn" to standard_quad_turn_points,
  quadratic turning points are now computed quite fast.

Mon 2 Apr 2007 : extended standard_quad_turn_points (and its _io) so
  that also quadratic turning points from a complex path are located
  and computed accurately via a plain bisection method.

Sat 31 Mar 2007 : added bisection method to seek turning points
  to standard_quad_turn_points, also the _io package changed.

Fri 30 Mar 2007 : created package standard_quad_turn_points + _io,
  from the code in ts_realcont.adb, which now locates turning points,
  interactively, for small examples like the circle...

Wed 28 Mar 2007 : started the creation of ts_realcont.adb to develop
  a real sweep for parameter continuation.

Tue 20 Mar 2007 : black_box_polynomials_continuations.adb did not contain
  a silent black_box_refine which caused different outputs in phc -b...

Sun 18 Feb 2007 : max step size in endgame should be 0.01 not 0.05.
  Changes continuation_parameters.a*.  This should decrease path crossings.

Thu 1 Feb 2007 : bug in black_box_poly_continuations.adb, instead of
  Random1 one was taken as the value for the gamma constant.

Mon 1 Jan 2007 : drivers_to_track_paths changed for more efficient
  solving of linear-product start systems.

Tue 17 Oct 2006 : changed drivers_to_track_paths.ads and .adb, moved
  maintrack to the Main directory for jumpstarting polyhedral homotopies.

Thu 28 Sep 2006 : changed drivers_for_poly_continuation.adb to help the
  user to set t to zero by suggesting default values in case t = 1.0.

Fri 23 Jun 2006 : changed drivers_for_poly_continuation.adb to fix a bug.

Thu 15 Jun 2006 : changed increment_and_fix_continuation.adb because of
  elimnation of standard_complex_solutions.get.

Wed 14 Jun 2006 : because of warning of AIX, changed valipoco.adb,
  standard_continuation_data.adb, drivers_to_track_paths.adb,
  parameter_homotopy_continuation.adb, drivers_for_poly_continuation.adb,
  black_box_poly_continuations.adb.

Thu 27 Apr 2006 : added another procedure for the next solution of
  a linear-product start system to drivers_to_track_paths.ads.

Tue 18 Apr 2006 : updated drivers_to_track_paths.ads and its adb.

Sun 26 Mar 2006 : updated standard_continuation_data_io with extra output
  when preliminary diagnostics of path; drivers_to_track_paths now supports
  jumpstarting from total degree and linear product start systems.

Sat 25 Mar 2006 : extended drivers_to_track_paths to work with start
  systems based on the total degree.

Fri 24 Mar 2006 : minor change in continuation_parameters_io.adb to allow
  for scanning of solution banner in the output of mpi2track.
  Discovered "black_polynomial_continuations" package to be obsolete.
  Modified black_box_poly_continuations.adb to write banner to file.

Sun 19 Feb 2006 : updated standard_continuation_data_io.ad* for lib_track.

Thu 16 Feb 2006 : created the file ts_track.adb to do simple path tracking
  with incremental read and write of the start solutions.  For extra output,
  created the package standard_continuation_data_io.  For maintrack, added
  the new package drivers_to_path_trackers.  Changed default for #corrector
  iterations back to 4 (from 3) because of decreased performance.

Thu 9 Feb 2006 : created the package pack_continuation_parameters, tested
  by ts_pactun, as needed to tune the parameters with the C interface;
  changed continuation_parameters.ads: 10^4 -> 10^8 for divergence along path

Tue 24 Jan 2006 : made some simplications in standard_predictors.adb to
  hunt for the bug in the Hermite predictor at the end of the paths.

Thu 19 Jan 2006 : updated continuation_parameters.ads, for the default
  predictor to be the Hermite cubic predictor, but it crashed!

Tue 17 Jan 2006 : updated continuation_parameters.ads, limiting the default
  number of corrector steps from 4 to 3.

Thu 24 Mar 2004 : extended process_io with some extra functions.

Mon 13 Dec 2004 : in Set_Homotopy_Parameters of black_box_poly_continuations,
  made sure the constant a was random1 to fix x*y = 2 and x+y = 3.

Thu 26 Aug 2004 : initialized err and dumbv in 
  increment_and_fix_continuation.adb to reduce errors reported by valgrind

Thu 12 Aug 2004 : updated ts_vlprs.adb to test pipelining and fixed
  ew estimate in {standard,multprec}_winding_numbers.adb.

Fri 6 Aug 2004 : moved the routines to extrapolate for winding number
  from ts_vlprs into {standard,multprec}_winding_numbers.

Thu 5 Aug 2004 : modified ts_vlprs to estimate winding number.

Wed 4 Aug 2004 : modified ts_vlprs and inserted standard_ to the
  packages vlprs_tables and vlprs_algorithm, created multprec versions.

Thu 29 Jul 2004 : added condition tables to the standard_root_refiners
  and multprec_root_refiners.

Tue 29 Jun 2004 : replaced deflation sequences by deflation trees,
  integrated *deflate_symbols in the standard_deflation_trees_io,
  updated drivers_to_deflate_singularities accordingly.

Fri 25 Jun 2004 : created standard and multprec deflation sequences,
  adjusted *deflate_singularties packages and ts_deflate.

Thu 24 Jun 2004 : changed drivers_to_deflate_singularities and ts_deflate
  to implement a deflation method operating on a list of solutions.

Sat 19 Jun 2004 : tolerance in Numerical_Rank in *newton_steps is parameter,
  changed also ts_deflate.adb accordingly.

Sat 8 May 2004 : added "Numerical_Rank" to *newton_steps, and updated
  ts_newton and ts_deflate accordingly.

Tue 4 May 2004 : wrote ts_deflate.adb, combining deflation with Newton,
  and also using multi-precision.

Sun 25 Apr 2004 : updated standard_deflate_singularities and ts_deflate.adb
  to have a variable number of multipliers in the deflated system.

Thu 22 Apr 2004 : updated ts_newton.adb, as a result of changes in
  standard_newton_steps and multprec_newton_steps.

Wed 3 Mar 2004 : in black_box_poly_continuations, changed the a := Random1
  to a := Create(1.0) to make sure no monodromy would happen.

Fri 27 Feb 2004 : changed bablpoco to have three file names as arguments
  and renamed black_polynomial_continuations into 
  black_box_poly_continuations; updated black_box routine.

Thu 18 Dec 2003 : fixed bug in Clear of standard_continuation_data.adb.

Fri 24 Oct 2003 : fixed bug in parameter_homotopy_continuation.adb,
  the example with the triangle in the book now works well

Fri 17 Oct 2003 : updated parameter_homotopy_continuation with a
  circular interpolate to force complex arithmetic

Sun 5 Oct 2003 : first try to have parameter continuation in phc -p,
  updated mainpoco.adb and systems_with_parameters.ads and .adb.

Wed 24 Sep 2003 : changed parameter_homotopy_continuation.adb

Tue 23 Sep 2003 : extra output with Hermite predictor removed from
  standard_path_trackers.adb and multprec_path_trackers.adb
  Also changed ts_deflate and standard_deflate_singularities.

Mon 22 Sep 2003 : ts_parcon.adb was splitted into systems_with_parameters,
  solutions_with_parameters, and parameter_homotopy_continuation.

Tue 8 Sep 2003 : continued development of standard_deflate_singularities
  and copied parts of ts_newton to ts_deflate.

Mon 1 Sep 2003 : started development of standard_deflate_singularities,
  standard_deflate_symbols and ts_deflate.

Fri 11 July 2003 : changed ts_numdif.adb and 
  standard_evaluation_machine.adb to deflate on a simple multiplicity.

************* most recent updates above this line *****************

Fri 23 May 2003 : standard_evaluation_machine,
  standard_numerical_derivatives, and ts_numdif are a first start
  to implement a Newton's method to treat singularities better.

Correctors :
  QR-decompostion in end of path following

Old_Correctors : no QR with conditioned
New_Correctors : QR when condition number is requested

Predictors : made the predictors for t visible
Dispatch_Predictors : "xt" = true pred for x+t, otherwise for t only
Continuation_Parameters.Tune : adapt reduction_*_factor with condition

Tue 10 August, 1999 : removed the infile from the parameter list of the
  Black_Box_Polynomial_Continuation that also has the start system as
  parameter, since the Tuning of the continuation parameters is not
  done from scanning the input file.
  Modification not copied to the public distribution directory.

Sat 14 August, 1999 : "put(outfile,p'last,p.all)" so that now also
  the dimension will appear when the system is put on outfile
  in the routine Black_Box_Polynomial_Continuation.
  Copied the file black_polynomial_continuations.ads and .adb to
  the public distribution directory.

Tue 16 May 2000 : ts_rootrefi.adb gave compilation error, because of missing
  alternative in case.  Changed the standard and multi-precision root refiners
  so that they print the correct start residual and the number of iterations.

Tue 30 May 2000 : implemented the method of Peter Kravanja, updated
  standard and multprec root refiners and ts_rootrefi.

Wed 31 May 2000 : modified specifications of existing routines in the
  root refiners, which led to changes in the following 12 files :
    Continuation/black_polynomial_continuations.adb
    Continuation/mainpoco.adb
    Root_Counts/Implift/mixed_homotopy_continuation.adb
    Root_Counts/Symmetry/mainsmvc.adb
    Schubert/pieri_continuation.adb
    Schubert/driver_for_sagbi_homotopies.adb
    Schubert/driver_for_pieri_homotopies.adb
    Schubert/driver_for_quantum_pieri.adb
    Components/component_scanners.adb
    Main/driver_for_root_refining.adb
    Main/bablvali.adb
    Main/mainvali.adb

Thu 8 June 2000 : started transition to multi-precision path tracking...
  created ts_preco, packages splitted into Standard and Multprec are
  Continuation_Data, Predictors, Dispatch_Predictors, Correctors,
  and Path_Trackers

Thu 15 June 2000 : ts_path works also for multi-precision numbers.

Sat 23 Sep 2000 : modified drivers_for_poly_continuation with an
  additional driver to tune the continuation  parameters.

Thu 5 Oct 2000 : added stop test to increment_and_fix_continuation

Wed 15 Nov 2000 : ts_path.adb used Homotopy instead of Standard_Homotopy.

Thu 30 Nov 2000 : directions_of_solution_paths.adb weeded out variables
  that were never assigned a value.

Thu 24 May 2001 : updated ts_path and drivers_for_poly_continuation to
  deal with multi-precision path tracking in a more user friendly way.
  Created new package multprec_incfix_continuation.

Fri 22 Mar 2002 : changed black_polynomial_continuations.adb so that
  the sols on return is not the list of end points, but the filtered
  list of true solutions, obtained after root refinement

Tue 30 July 2002 : scaling -> standard_scaling changes 
   black_polynomial_continuations.adb

Wed 25 Dec 2002 : changed continuation_data, adding lists of arrays
  of solu_info to deal with segmentation faults when tracking huge
  lists of solution paths.  Changed increment_and_fix_continuation
  which should to run with these lists of arrays of solu_info.

Fri 16 May 2003 : mainpoco.adb, cosmetic change: 
  "Reading target polynomial system..."
