ginac.git
20 years ago- typos, typos, typos...
Richard Kreckel [Fri, 4 Feb 2000 21:22:10 +0000 (21:22 +0000)]
- typos, typos, typos...

20 years ago- ginsh now includes "ginac.h" when compiled as part of the GiNaC package,
Christian Bauer [Fri, 4 Feb 2000 18:03:55 +0000 (18:03 +0000)]
- ginsh now includes "ginac.h" when compiled as part of the GiNaC package,
  and <ginac/ginac.h> otherwise

20 years agobug fix for output of interactive expression in ginaccint
Alexander Frink [Thu, 3 Feb 2000 00:56:21 +0000 (00:56 +0000)]
bug fix for output of interactive expression in ginaccint

20 years agobug fix for output of interactive expression
Alexander Frink [Thu, 3 Feb 2000 00:56:08 +0000 (00:56 +0000)]
bug fix for output of interactive expression

20 years ago- Change triggered by newer automake.
Richard Kreckel [Wed, 2 Feb 2000 23:09:31 +0000 (23:09 +0000)]
- Change triggered by newer automake.

20 years ago- Extended manpage.
Richard Kreckel [Wed, 2 Feb 2000 23:09:08 +0000 (23:09 +0000)]
- Extended manpage.
- Made make actually work.  :-)

20 years ago- use namespace GiNaC only if NO_NAMESPACE_GINAC wasn't defined.
Richard Kreckel [Wed, 2 Feb 2000 23:08:10 +0000 (23:08 +0000)]
- use namespace GiNaC only if NO_NAMESPACE_GINAC wasn't defined.

20 years ago- Make diff() care for evaluating stuff.
Richard Kreckel [Wed, 2 Feb 2000 21:07:07 +0000 (21:07 +0000)]
- Make diff() care for evaluating stuff.

20 years ago- Several corrections in section Extending GiNaC / Symbolic Functions.
Richard Kreckel [Wed, 2 Feb 2000 21:06:33 +0000 (21:06 +0000)]
- Several corrections in section Extending GiNaC / Symbolic Functions.

20 years ago- diff() is now only defined on classes ex and basic, where it handles
Richard Kreckel [Wed, 2 Feb 2000 18:49:29 +0000 (18:49 +0000)]
- diff() is now only defined on classes ex and basic, where it handles
  higher derivatives.  derivative() is now the method to be implemented
  by the user that does single derivatives, it is protected.
- Implemented some very clumpsy automatic generation of ginaccint.
  It is ugly because we have to compile libginac without namespace GiNaC.
- Fixed some wrong logic in numeric::power().

20 years ago- introduced function void greeting(void).
Richard Kreckel [Mon, 31 Jan 2000 22:50:12 +0000 (22:50 +0000)]
- introduced function void greeting(void).

20 years ago- archive file format is a little more compact (stores property name and type
Christian Bauer [Mon, 31 Jan 2000 21:50:12 +0000 (21:50 +0000)]
- archive file format is a little more compact (stores property name and type
  as one unsigned value)
- moved #include "config.h" from lortensor.h to lortensor.cpp (none of the
  public headers must include config.h)

20 years ago- fixed bug in multiply_lcm() (see paranoia_check9)
Christian Bauer [Mon, 31 Jan 2000 21:32:56 +0000 (21:32 +0000)]
- fixed bug in multiply_lcm() (see paranoia_check9)

20 years ago- sources now include "ginac.h" instead of <ginac/ginac.h>, changed Makefile.am
Christian Bauer [Mon, 31 Jan 2000 21:32:23 +0000 (21:32 +0000)]
- sources now include "ginac.h" instead of <ginac/ginac.h>, changed Makefile.am
  to add ../ginac to include path

20 years ago- updated Makefiles
Christian Bauer [Mon, 31 Jan 2000 21:31:46 +0000 (21:31 +0000)]
- updated Makefiles

20 years ago- Renamed flag NO_GINAC_NAMESPACE to NO_NAMESPACE_GINAC because of m4.
Richard Kreckel [Sun, 30 Jan 2000 23:52:53 +0000 (23:52 +0000)]
- Renamed flag NO_GINAC_NAMESPACE to NO_NAMESPACE_GINAC because of m4.
- Made configure check for cint and makecint and call makecint to dump out
  cint/Makefile.cint, which will later be called by cint/Makefile.  See
  configure option --with-cint.
- Added some documentation about GiNaC-cint (e.g. a manpage).
- All dummies in subdir cint/ are created by cint/dummies.pl now, removed
  older stuff.

20 years agogenheader.pl: create a header file 'cint_workaround.h' to fix the broken
Alexander Frink [Fri, 28 Jan 2000 18:26:41 +0000 (18:26 +0000)]
genheader.pl: create a header file 'cint_workaround.h' to fix the broken
cint overloading for C math library functions

20 years ago- moved configure options behind libtool so they appear at the end of the list
Christian Bauer [Fri, 28 Jan 2000 17:46:11 +0000 (17:46 +0000)]
- moved configure options behind libtool so they appear at the end of the list
  in "configure --help"

20 years ago- final tweaks to spec file
Christian Bauer [Fri, 28 Jan 2000 15:34:01 +0000 (15:34 +0000)]
- final tweaks to spec file

20 years ago- building GiNaC in a separate directory now works
Christian Bauer [Thu, 27 Jan 2000 18:26:09 +0000 (18:26 +0000)]
- building GiNaC in a separate directory now works

20 years ago- man page is now put in distribution
Christian Bauer [Wed, 26 Jan 2000 20:44:27 +0000 (20:44 +0000)]
- man page is now put in distribution

20 years ago- updated dates
Christian Bauer [Wed, 26 Jan 2000 20:44:03 +0000 (20:44 +0000)]
- updated dates

20 years ago- updated Makefiles
Christian Bauer [Wed, 26 Jan 2000 20:43:53 +0000 (20:43 +0000)]
- updated Makefiles

20 years ago- RPM packages split into user and devel packages
Christian Bauer [Wed, 26 Jan 2000 20:43:39 +0000 (20:43 +0000)]
- RPM packages split into user and devel packages
- added missing stuff to RPM packages

20 years ago- gcd(): cofactor computation is faster in partially factorized case
Christian Bauer [Wed, 26 Jan 2000 17:19:05 +0000 (17:19 +0000)]
- gcd(): cofactor computation is faster in partially factorized case

20 years ago- lcm_of_coefficients() works for unexpanded polynomials
Christian Bauer [Tue, 25 Jan 2000 22:35:47 +0000 (22:35 +0000)]
- lcm_of_coefficients() works for unexpanded polynomials
- frac_cancel() works for unexpanded polynomials
- gcd() can handle partially-factored input polynomials without expanding them

20 years ago- checks adapted to new GCD behavior for partly-factored input polynomials
Christian Bauer [Tue, 25 Jan 2000 22:31:49 +0000 (22:31 +0000)]
- checks adapted to new GCD behavior for partly-factored input polynomials

20 years ago- mentioned archiving
Christian Bauer [Tue, 25 Jan 2000 22:31:21 +0000 (22:31 +0000)]
- mentioned archiving

20 years ago- Updated section about CLN to current status (need 1.0.3)
Richard Kreckel [Tue, 25 Jan 2000 19:58:44 +0000 (19:58 +0000)]
- Updated section about CLN to current status (need 1.0.3)

20 years ago- Added some comments for doxygen.
Richard Kreckel [Mon, 24 Jan 2000 22:20:04 +0000 (22:20 +0000)]
- Added some comments for doxygen.

20 years ago- removed superfluous ex ser_numer; in gamma_series()
Richard Kreckel [Mon, 24 Jan 2000 22:19:37 +0000 (22:19 +0000)]
- removed superfluous ex ser_numer; in gamma_series()

20 years ago- Corrected a couple of typos
Richard Kreckel [Mon, 24 Jan 2000 22:17:13 +0000 (22:17 +0000)]
- Corrected a couple of typos

20 years ago- added missing method matrix::op()
Richard Kreckel [Mon, 24 Jan 2000 22:15:31 +0000 (22:15 +0000)]
- added missing method matrix::op()

20 years ago- Added Fibonacci numbers for integers
Richard Kreckel [Mon, 24 Jan 2000 22:12:31 +0000 (22:12 +0000)]
- Added Fibonacci numbers for integers
- many functions now return const to trap errors like fibonacci(1)=10;

20 years ago- killed old add::print
Richard Kreckel [Mon, 24 Jan 2000 22:09:46 +0000 (22:09 +0000)]
- killed old add::print

20 years ago- archiving implemented in lorentzidx
Christian Bauer [Mon, 24 Jan 2000 21:28:22 +0000 (21:28 +0000)]
- archiving implemented in lorentzidx

20 years ago- Changed the name of cache-value ac_cv_header_CLN_cln_h to
Richard Kreckel [Mon, 24 Jan 2000 21:24:00 +0000 (21:24 +0000)]
- Changed the name of cache-value ac_cv_header_CLN_cln_h to
  ac_cv_header_cln_cln_h to reflect new pathname cln/cln.h.

20 years ago- renamed archive::dump() to archive::printraw() for consistency with the
Christian Bauer [Mon, 24 Jan 2000 20:53:56 +0000 (20:53 +0000)]
- renamed archive::dump() to archive::printraw() for consistency with the
  other classes
- implemented archiving for the following classes: color, idx, coloridx,
  isospin
- non-crational numbers are now stored in an integer-decoded format in archives

20 years ago- Macro GINAC_CHECK_LIBCLN only checks if doublefactorial is available now.
Richard Kreckel [Mon, 24 Jan 2000 15:40:42 +0000 (15:40 +0000)]
- Macro GINAC_CHECK_LIBCLN only checks if doublefactorial is available now.
  If it is, the version is new enough and we can omit the second sanity
  test which makes that configure step twice as fast.

20 years ago- Fixed old class series stuff to class pseries
Richard Kreckel [Fri, 21 Jan 2000 19:48:09 +0000 (19:48 +0000)]
- Fixed old class series stuff to class pseries
- Added little overview of classes under hierarchy tree

20 years ago- changed all instances of "foo const &/*" to "const foo &/*"
Christian Bauer [Fri, 21 Jan 2000 18:20:26 +0000 (18:20 +0000)]
- changed all instances of "foo const &/*" to "const foo &/*"
- changed function arguments like "int const" to a simple "int"

20 years ago- implemented nops() and op() for pseries objects
Christian Bauer [Fri, 21 Jan 2000 17:51:26 +0000 (17:51 +0000)]
- implemented nops() and op() for pseries objects

20 years ago- Use newer fig2dev which supports eps (old ones are broken now!)
Richard Kreckel [Fri, 21 Jan 2000 13:58:06 +0000 (13:58 +0000)]
- Use newer fig2dev which supports eps (old ones are broken now!)

20 years ago- Little addition about something unmentioned but ovious: x^ax^b=x^(a+b)
Richard Kreckel [Fri, 21 Jan 2000 13:55:04 +0000 (13:55 +0000)]
- Little addition about something unmentioned but ovious: x^ax^b=x^(a+b)

20 years ago- changed old-style power() to new-style pow()
Richard Kreckel [Thu, 20 Jan 2000 00:14:33 +0000 (00:14 +0000)]
- changed old-style power() to new-style pow()

20 years ago- Changed policy: look for cln/cln.h instead of CLN/cln.h, reflecting an
Richard Kreckel [Wed, 19 Jan 2000 23:51:21 +0000 (23:51 +0000)]
- Changed policy: look for cln/cln.h instead of CLN/cln.h, reflecting an
  agreement with Bruno Haible.
- Don't #include <cln/cln.h> but only those headers necessary for GiNaC.

20 years ago- Changed policy: look for cln/cln.h instead of CLN/cln.h, reflecting an
Richard Kreckel [Wed, 19 Jan 2000 23:50:48 +0000 (23:50 +0000)]
- Changed policy: look for cln/cln.h instead of CLN/cln.h, reflecting an
  agreement with Bruno Haible.

20 years ago- change triggered by newer autoconf version
Richard Kreckel [Wed, 19 Jan 2000 23:48:55 +0000 (23:48 +0000)]
- change triggered by newer autoconf version

20 years ago- updated class pseries
Richard Kreckel [Wed, 19 Jan 2000 23:48:01 +0000 (23:48 +0000)]
- updated class pseries

20 years ago- change triggered by newer automake version
Richard Kreckel [Wed, 19 Jan 2000 23:44:48 +0000 (23:44 +0000)]
- change triggered by newer automake version

20 years ago- created
Christian Bauer [Wed, 19 Jan 2000 22:47:35 +0000 (22:47 +0000)]
- created

20 years ago- added tokens "read" and "write" (in preparation for archiving)
Christian Bauer [Wed, 19 Jan 2000 22:47:28 +0000 (22:47 +0000)]
- added tokens "read" and "write" (in preparation for archiving)

20 years ago- implemented global class registry (for class basic and derived classes)
Christian Bauer [Wed, 19 Jan 2000 22:46:56 +0000 (22:46 +0000)]
- implemented global class registry (for class basic and derived classes)
- implemented archiving of expressions (except for class idx, and all classes
  derived from idx and indexed)
- added series() wrapper function
- class series renamed to pseries to avoid name clashes with global series()
  wrapper function, ex::series(), and basic::series()
- corrected the series expansion of single symbols with respect to the order
  term
- calling subs() on a series didn't work

20 years ago- added a chapter about archiving of expressions
Christian Bauer [Wed, 19 Jan 2000 22:43:17 +0000 (22:43 +0000)]
- added a chapter about archiving of expressions
- updated makefiles

20 years ago- updated makefiles
Christian Bauer [Wed, 19 Jan 2000 22:42:47 +0000 (22:42 +0000)]
- updated makefiles

20 years ago- updated for class series->class pseries
Christian Bauer [Wed, 19 Jan 2000 22:42:36 +0000 (22:42 +0000)]
- updated for class series->class pseries
- added one additional check for the behavior of the order term during the
  series expansion of a simple polynomial

20 years ago- updated configure script and makefiles
Christian Bauer [Wed, 19 Jan 2000 22:41:40 +0000 (22:41 +0000)]
- updated configure script and makefiles

20 years agoModification in output of last returned expression (for ginaccint)
Alexander Frink [Mon, 17 Jan 2000 19:49:03 +0000 (19:49 +0000)]
Modification in output of last returned expression (for ginaccint)

20 years agoModification in output of last returned expression
Alexander Frink [Mon, 17 Jan 2000 19:48:53 +0000 (19:48 +0000)]
Modification in output of last returned expression

20 years ago- Importing new files for packaging Debian GNU/Linux packages.
Richard Kreckel [Mon, 17 Jan 2000 19:45:57 +0000 (19:45 +0000)]
- Importing new files for packaging Debian GNU/Linux packages.

20 years ago- doublefactorial now falls back directly to CLN, which is much faster.
Richard Kreckel [Sat, 15 Jan 2000 21:20:20 +0000 (21:20 +0000)]
- doublefactorial now falls back directly to CLN, which is much faster.

20 years ago- made nops() return unsigned instead of int
Richard Kreckel [Tue, 11 Jan 2000 22:49:07 +0000 (22:49 +0000)]
- made nops() return unsigned instead of int

20 years ago- changed the fucking dates i had forgotten the last time...
Richard Kreckel [Tue, 11 Jan 2000 22:28:31 +0000 (22:28 +0000)]
- changed the fucking dates i had forgotten the last time...

20 years ago- made nops() return unsigned instead of int
Richard Kreckel [Tue, 11 Jan 2000 22:25:18 +0000 (22:25 +0000)]
- made nops() return unsigned instead of int
- changed dates in copyright notices
- added psi and zeta to online help in ginsh

20 years ago- made nops() return unsigned instead of int
Richard Kreckel [Tue, 11 Jan 2000 22:25:05 +0000 (22:25 +0000)]
- made nops() return unsigned instead of int
- changed dates in copyright notices

20 years ago- changed dates in copyright notices
Richard Kreckel [Tue, 11 Jan 2000 22:24:49 +0000 (22:24 +0000)]
- changed dates in copyright notices

20 years ago- now compiles with readline V4 headers
Christian Bauer [Mon, 10 Jan 2000 21:00:21 +0000 (21:00 +0000)]
- now compiles with readline V4 headers

20 years ago- minor updates.
welzel [Mon, 10 Jan 2000 19:03:18 +0000 (19:03 +0000)]
- minor updates.

20 years ago- added simple checks of manipulations with lortensor objects.
welzel [Mon, 10 Jan 2000 19:02:27 +0000 (19:02 +0000)]
- added simple checks of manipulations with lortensor objects.

20 years ago- speedup by declaring x_pt and so on const
Richard Kreckel [Sat, 25 Dec 1999 21:09:06 +0000 (21:09 +0000)]
- speedup by declaring x_pt and so on const
- added some code for series expansion of beta function (untested!)

20 years ago- some indentation fixes
Richard Kreckel [Sat, 25 Dec 1999 21:08:16 +0000 (21:08 +0000)]
- some indentation fixes

20 years ago- polygamma functions (psi functions) know about half-integer arguments
Richard Kreckel [Wed, 22 Dec 1999 17:51:32 +0000 (17:51 +0000)]
- polygamma functions (psi functions) know about half-integer arguments
  now, which is more than Mathematica's PolyGamma[n,x] provides.  :-)

20 years ago- matrix row/col made unsigned
Christian Bauer [Tue, 21 Dec 1999 17:25:10 +0000 (17:25 +0000)]
- matrix row/col made unsigned

20 years ago- introduced checks that test the new logic of sin, cos and tan to control
Richard Kreckel [Mon, 20 Dec 1999 22:14:33 +0000 (22:14 +0000)]
- introduced checks that test the new logic of sin, cos and tan to control
  transformations like sin(Pi/12) -> sqrt(6)/4*(1-sqrt(3)/3)

20 years ago- added a abs() function. haven't you been longing for this?
Richard Kreckel [Mon, 20 Dec 1999 22:12:02 +0000 (22:12 +0000)]
- added a abs() function. haven't you been longing for this?

20 years ago- more logic on the trigonometric function stuff.
Richard Kreckel [Mon, 20 Dec 1999 20:12:36 +0000 (20:12 +0000)]
- more logic on the trigonometric function stuff.
- changed several occurences of numeric const & to const numeric &, which
  is the same, but doxygen kept being confused because declarations
  differed from implementations.

20 years ago- Added a clarification about branch cuts.
Richard Kreckel [Mon, 20 Dec 1999 20:06:53 +0000 (20:06 +0000)]
- Added a clarification about branch cuts.

20 years ago- Changed a few switches.
Richard Kreckel [Mon, 20 Dec 1999 20:06:35 +0000 (20:06 +0000)]
- Changed a few switches.

20 years ago- more logic on the trigonometric function stuff.
Richard Kreckel [Mon, 20 Dec 1999 20:06:12 +0000 (20:06 +0000)]
- more logic on the trigonometric function stuff.
- changed several occurences of numeric const & to const numeric &, which
  is the same, but doxygen kept being confused because declarations
  differed from implementations.

20 years ago- Banned exZERO(), exONE(), exMINUSHALF() and all this from the interface.
Richard Kreckel [Fri, 17 Dec 1999 19:58:25 +0000 (19:58 +0000)]
- Banned exZERO(), exONE(), exMINUSHALF() and all this from the interface.
  There is now _ex1() meaning 1, _ex_1() meaning -1, _ex1_2() meaning 1/2
  and so on defined in utils.h and implemented in utils.cpp.  Feel free
  to extend them as it pleases you but use them inside the library only.
- Added more evaluations of trigonometric functions
- Added series expansions for psi(x) and psi(n,x) at all their poles

20 years agolortensor class
welzel [Fri, 17 Dec 1999 17:44:17 +0000 (17:44 +0000)]
lortensor class

20 years ago- Corrected some minor texinfo flaws in tutorial
Richard Kreckel [Tue, 14 Dec 1999 21:35:41 +0000 (21:35 +0000)]
- Corrected some minor texinfo flaws in tutorial

20 years agofirst release of GiNaCcint
Alexander Frink [Tue, 14 Dec 1999 21:23:29 +0000 (21:23 +0000)]
first release of GiNaCcint

20 years agosome changes to allow GiNaC to cooperate with cint:
Alexander Frink [Tue, 14 Dec 1999 20:41:18 +0000 (20:41 +0000)]
some changes to allow GiNaC to cooperate with cint:

matrix.h/.cpp: vector<ex> replaced with exvector where
               typedef vector<ex,malloc_alloc> exvector

constant.h/.cpp: typedef for evalffunction

color.h/.cpp: replaced vector<exvector> with exvectorvector where
              typedef vector<exvector,malloc_alloc> exvectorvector

ex.h: OBSCURE_CINT_HACK defines around some weird code. Never define
      OBSCURE_CINT_HACK when compiling libginac

20 years ago- Brought to the 0.4.1-level.
Richard Kreckel [Mon, 13 Dec 1999 19:31:34 +0000 (19:31 +0000)]
- Brought to the 0.4.1-level.

20 years ago- Run automake again on squark, as we had agreed earlier.
Richard Kreckel [Mon, 13 Dec 1999 18:44:25 +0000 (18:44 +0000)]
- Run automake again on squark, as we had agreed earlier.

20 years ago- introduced info_flags::cinteger, info_flags::crational,
Richard Kreckel [Mon, 13 Dec 1999 18:43:11 +0000 (18:43 +0000)]
- introduced info_flags::cinteger, info_flags::crational,
  info_flags::cinteger_polynomial, info_flags::crational_polynomial with
  intuitive behaviour.
- extended documentation
- made things like evlaf(zeta(3)) work
- killed several bugs

20 years ago- changed behaviour of numeric::is_rational() and added numeric::is_cinteger()
Richard Kreckel [Fri, 10 Dec 1999 20:23:46 +0000 (20:23 +0000)]
- changed behaviour of numeric::is_rational() and added numeric::is_cinteger()
  and numeric::is_crational() which work for complex numbers now
- added default ctor for constant to please CInt
- documentation reflects the changes for upcoming 0.4.1

20 years ago- Introduced exception do_taylor to signal Taylor expansion is ok for series
Richard Kreckel [Fri, 10 Dec 1999 14:59:18 +0000 (14:59 +0000)]
- Introduced exception do_taylor to signal Taylor expansion is ok for series
- Finished a clean implementation of gamma function's series expansion
- Bumped up version number to 1.4.1 to reflect the changes and the
  soon-to-come second prerelease.

20 years ago- added the beta function to GiNaC
Richard Kreckel [Thu, 2 Dec 1999 22:28:00 +0000 (22:28 +0000)]
- added the beta function to GiNaC
- threw out ginsh's beta function
- added a check for first polygamma function

20 years agoginsh did not compile
Alexander Frink [Thu, 2 Dec 1999 18:55:08 +0000 (18:55 +0000)]
ginsh did not compile

20 years agofunction.pl did not compile
Alexander Frink [Thu, 2 Dec 1999 18:49:37 +0000 (18:49 +0000)]
function.pl did not compile

20 years ago#ifndef around namespace GiNaC { }
Alexander Frink [Thu, 2 Dec 1999 18:44:55 +0000 (18:44 +0000)]
#ifndef around namespace GiNaC { }

20 years ago- added prefix and postfix increment and decrement operators for class numeric
Richard Kreckel [Wed, 1 Dec 1999 21:13:01 +0000 (21:13 +0000)]
- added prefix and postfix increment and decrement operators for class numeric
- made all function_index const
- added overloaded psi(x) and psi(n,x) with some sensible behaviour

20 years ago- added 'ginac-config' script and 'ginac.m4' macros to help making packages
Christian Bauer [Wed, 1 Dec 1999 19:03:37 +0000 (19:03 +0000)]
- added 'ginac-config' script and 'ginac.m4' macros to help making packages
  that use GiNaC
- installation does no longer fail when HTML docs were disabled by
  configuration

20 years ago- changed mul::print() to behave similar to add::print()
Richard Kreckel [Tue, 30 Nov 1999 22:06:14 +0000 (22:06 +0000)]
- changed mul::print() to behave similar to add::print()

20 years ago- changed function::diff() to be more tolerant by checking first if the
Richard Kreckel [Mon, 29 Nov 1999 22:50:42 +0000 (22:50 +0000)]
- changed function::diff() to be more tolerant by checking first if the
  nth argument when differentiated is non-zero and *then* building the sum.
- added support for overloaded polygamma functions psi(x) and psi(n,x).
- changed return code of atan2_diff to be somewhat simpler and adjusted
  check/differentiation.cpp to account for this.

20 years ago- clarified timings
Richard Kreckel [Mon, 29 Nov 1999 22:46:23 +0000 (22:46 +0000)]
- clarified timings

20 years ago- alignment fix. :-)
Richard Kreckel [Mon, 29 Nov 1999 22:46:04 +0000 (22:46 +0000)]
- alignment fix.  :-)

20 years ago- corrected an error due to my ignorance WRT diff_param
Richard Kreckel [Mon, 29 Nov 1999 22:23:36 +0000 (22:23 +0000)]
- corrected an error due to my ignorance WRT diff_param