ChangeLog

30 Aug 2022:

- Fixed RegisterAbbr issue on abbreviations of the form DiracChain[.]*Eps[.].

- Added way to implement crossing (beta).


15 Jul 2022:

- Worked around yet another Mathematica segfault due to RCSub = Simplify.


29 Mar 2022:

- Indices in IndexSums etc. now get renumbered with similar index names
  (no Ind1, Ind2, ...).


5 Jan 2022:

- DiagramType now takes powers of denominators properly into account.


29 Nov 2021:

- Added ReadFormWrap for better debugging of Mathematica-FORM handshake.


7 Jun 2021:

- Added VectorMoms option to CalcFeynAmp to return loop integrals
  with vectors instead of invariants as arguments.


18 Oct 2019:

- Moved nf into SM common blocks.

- Minor cosmetic changes.


7 Oct 2019:

- Automated generation of model initialization code, now takes care of 
  FeynRules' .pars files.

- Warning: FormCalc does not, of course, guarantee that the numerical 
  parameters transmitted by FeynRules are compatible with the chosen 
  renormalization scheme. 


2 Sep 2019:

- Fixed gfortran hiccup with integer*8 in *MomDecode.F.

- Upgraded to latest FORM version.


10 Jun 2019:

- Fixed issue with IndexSum on fermion chains.


12 Apr 2019:

- Computation of generic fermion amplitudes fixed (unresolved SpinorType).

- Improved partial fractioning of loop denominators in FORM.

- Removed spurious IGram instances for PaVeReduce -> LoopTools.

- Added $Dminus4MaxPower (= 1 for one-loop).


10 Apr 2019:

- Fixed a few bugs introduced in the last release, most notably one
  in the F90 rework of ODESolve.F.


4 Apr 2019:

- Added better simplification of (numerical) roots, e.g. Sqrt[2],
  in FORM.


29 Mar 2019:

- Overhauled large parts of the drivers, put Fortran vector syntax
  in many places.

- Split xsection.F's TotalXS into IterateXS and (new) TotalXS, the latter
  just for the computation of the XS without looping over phase space.

- Added latest FORM executables.


8 Mar 2019:

- Moved partial fractioning of loop denominators...
	loop[a___, Den[p_, m1_], b___, Den[p_, m2_], c___] :=
	  loop[a, Den[p, m1] - Den[p, m2], b, c]/(m1 - m2) /;
	  m1 =!= m2
  ... to CalcFeynAmp.frm since m1 =!= m2 cannot reliably be evaluated
  before IndexSimplify.


18 Feb 2019:

- Introduced $KeepAbbr = False to not save the Subexpr/Abbr in Keep.
  Take care: expressions saved this way cannot not in general be fully
  reconstructed later.


29 Jan 2019:

- More quirks in the Generic name mangling removed.


28 Jan 2019:

- Fixed name issue with Generic object 'MetricTensor'.

- Improved persistent names for Generic objects.


21 Jan 2019:

- Fixed bug in CalcFeynAmp for explicit Dirac indices.

- Version number now available as
  $FormCalc = {major, minor},
  $FormCalcVersionNumber = major.minor
  $FormCalcVersion = "FormCalc major.minor (rev date)"


29 Dec 2018:

- Fixed issue with Mass[..., Loop] in Generic diagram simplification.


24 Dec 2018:

- Fixed compatibility issue with Mathematica version 9 and below.


4 Dec 2018:

- Introduced persistent names for Generic objects (couplings, masses etc.).


18 Sep 2018:

- ColourSimplify now observes SumOver for Colour and Gluon indices.
  This is needed for 'hand' calculations only since amplitudes from
  CalcFeynAmp never have explicit Colour/Gluon SumOvers.


3 Sep 2018:

- Fixed trivial error (duplicate variable in Block) in PolarizationSum.


17 Aug 2018:

- Type of dup variables in FF modules dynamically determined
  (no longer always HelType, wrong if HelicityME used).

- More settings for Type, TmpType, IndexType options.


4 Aug 2018:

- Fixed wrong detection of Dirac chains in the counting of the
  d.o.f. in the helicity loop of WriteSquaredME.


3 Aug 2018:

- Removed substituted vars from CodeExpr in Subst[Simple]Abbr.


2 Aug 2018:

- Improved Subst[Simple]Abbr: now also looks inside IndexIf.


30 Jul 2018:

- Fixed indexing problem in OptimizeAbbr.


27 Jul 2018:

- Added VarSort and Prio for better control of sort order of lists
  of definitions (var -> val).


23 Jul 2018:

- Fixed bug in WeylME.


19 Jul 2018:

- (Again) disabled SIMD detection for gfortran.


16 Jul 2018:

- Fixed two small issues with library detection in configure.


11 Jul 2018:

- SubstAbbr, SubstSimpleAbbr accept a further optional 'deny' argument.
  Variables are substituted only if they do not match 'deny'.

- Added HoldCode[expr] which preserves expr until write-out to
  Fortran or C.  Similar to HoldForm except that argument is
  not held unevaluated.


7 Jul 2018:

- Extended Abbreviate[expr, func] so that func can return True and False
  (as before) and also a subexpression of its argument, for which the
  abbreviation is then introduced.

- New function FindDeps finds all variables in a list of definitions
  whose rhs directly or indirectly depends on a given pattern.


27 Jun 2018:

- Added f @ m notation for {Mass,Field,Width}RC, as in:
  MassRC[S[3] @ MA0tree].


16 Apr 2018:

- Fixed `mysterious' problem in simplification code in PolarizationSum.frm
  introduced by last release.

- Removed bottleneck behavior in TermCollect.


19 Mar 2018:

- More efficient collecting of terms in PolarizationSum.


15 Mar 2018:

- PrepareExpr now generates `individualized' tmp and dup var names,
  e.g. C0tmp1 for an abbreviated C0.  Can choose prefixes with
  $TmpPrefix and $DupPrefix, too.


7 Mar 2018:

- Added DoDim (like Dim) which collects the index dimensions
  gathered from SumOvers during amplitude evaluation.

- Manual assignments to Dim make for correct array decls,
  assignments to DoDim also generate loops.


5 Mar 2018:

- Fixed problem with DenyNoExp and SUN objects.

- (Hopefully) improved IndexSimplify for terms containing IndexDeltas.


21 Feb 2018:

- Fixed SquaredME problem with amplitudes containing no Mat.


20 Feb 2018:

- Fixed SquaredME problem with threading of Mat.


12 Feb 2018:

- (Hopefully) fixed the dreaded NumberMalloc problem in FORM.


10 Feb 2018:

- Added missing end-braces in C's $CodeElse, $CodeElseIf.

- Added Declarations option to PrepareExpr.


26 Jan 2018:

- Improvements to several scripts in drivers/tools, especially pnuglot.

- Fixed bug in WriteExpr for debug lines in DoLoops.


5 Jan 2018:

- Added $RCRes (like $RCTop, $RCIns, $RCAmp) for better diagnostics
  in [D]SelfEnergy, *RC, RenConst, etc.

- Added SEHook to intercept final calculation of a [D]SelfEnergy and
  replacement of K2 by m2.

- Generalized SubstSimpleAbbr to SubstAbbr, of which SubstSimpleAbbr
  is now only a special case.

- Added Enum and ClearEnum which allow determination of array bounds
  with named constants in the code generation.

- Renamed OptPaint to PaintSE (function corresponding to $PaintSE).

- Added PutSE which saves the four components of the calculation of
  a RenConst, {$RCTop, $RCIns, $RCAmp, $RCRes}, in files, with
  filename conventions analogous to PaintSE.  Variable $PutSE
  sets the default path/prefix, analogous to $PaintSE.

- Added ToVars function which introduces variables for all
  subexpressions matching a pattern.

- Added MakeTmp option for PrepareExpr, which specifies a function
  to introduce user-defined temporary variables, e.g. using ToVars.

- Added DeclIf option for WriteExpr.  For DeclIf -> "var" the
  output is structured like
	#ifndef var
	#define var
	[declarations]
	#else
	[code]
	#endif
   i.e. in a language that requires separation of declarations
   and code (like Fortran) the same file can be included once
   in the declarations section and once in the code section.

- Added the NoDebug wrapper for statements given to PrepareExpr,
  i.e. a statement of the form NoDebug[var -> expr] does not
  produce a debug line.

- Added Unabbr[expr, !patt] to not expand abbreviations and
  subexpressions inside objects matching patt.

- Fixed the (rather special) case where PolarizationSum is used to
  square an amplitude that does not contain matrix elements.
  PolarizationSum now retains Mat[...] not substituted by the user
  (though a warning is issued if the Mat contain polarization
  vectors).

- Improved tools/USfSimplify.m.


13 Jun 17:

- Fixed conjugation problem in SquaredME.

- Updated FeynHiggs interface to 2.13.


13 Apr 17:

- IMPORTANT CHANGE: HelicityME, ColourME, WeylME, SquaredME now take
  arguments in the order "Mtree, Mloop".  This is now (finally) the
  same as with WriteSquaredME but different than before.

- Abbr[] and Subexpr[] now return lists with 'patternized' l.h.s. so
  that they can also be used for replacing inside Mathematica.

- New functions Creep, ExprHeads, ExprParts.

- Enhancements to ApplyUnitarity, Keep.

- Internally restructured HelicityME, PolarizedSum.


8 Mar 17:

- Fixed bug in OptimizeAbbr (typo: forgotten comma).

- Restricted substitution <-6,mu,nu> <-7,mu,nu> = 0 to Dirac
  chains with exactly two Lorentz indices.


23 Feb 17:

- Fixed another bug from returning Abbr/Subexpr with patterns
  in issuing debug statements.


10 Feb 17:

- Rearranged SquaredME (and the FORM code for PolarizationSum)
  to reflect the general structure of the squared amplitude,
  i.e. the squared amplitude is written in terms of form factors
  FF[.] or FFC[.], which are given by substitution rules.

- Improved Keep function, will store just the given expression
  if it does not depend on any Abbr[] or Subexpr[], else writes
  out AbbrExpr[expr, abbr, subexpr].

- Independent of the SetLanguage choice WriteRenConst now generates
  two sets of declarations, for Fortran (.h.F) and C (.h.c) so that
  Fortran drivers can conveniently be combined with generated
  C code (and vice versa).


24 Jan 17:

- Fixed several bugs in RegisterAbbr and RegisterSubexpr
  (introduced when Abbr/Subexpr were returned with patterns).

- Enhanced the Keep function: will store/load expression only
  if no Abbr and Subexpr are defined.

- New function ExprHeads returns all non-system symbols and heads.

- New function ExprParts returns all subexpressions of expr
  which contain given symbols and functions.

- New options RowIndex, ColumnIndex, ComplementFunction for
  ApplyUnitarity.

- New function Creep which applies a function only to
  subexpressions containing given patterns.


1 Aug 16:

- Abbr[] and Subexpr[] now return lists with 'patternized' l.h.s.
  so that they can also be used for replacing inside Mathematica.


9 Jun 16:

- Fixed non-removal of internal dot[.,.] functions in FORM code.


8 Jun 16:

- Automated detection of unpolarized Dirac fermions in the
  Fortran code, i.e. got rid of manually setting
  UNPOLARIZED_DIRAC_FERMIONS in process.h altogether.
  For this to work the definitions of Hel must be present during
  WriteSquaredME, however.


7 Jun 16:

- Corrected averaging factor for fermions computed with the
  HelicityME method.

- Added incoming/outgoing information to 'Compose' preprocessor
  function in specs.h.

- Fixed empty variable declaration in FORM for Invariants -> False.


4 Jun 16:

- Fixed issue with HelNaN, which was not declared in a non-vectorized
  environment.


3 Jun 16:

- Moved FORM procedures common to CalcFeynAmp, HelicityME, 
  PolarizationSum to Common.frm (with slight adaptations).

- In FormCalc.m accordingly moved code related to momentum 
  conservation (including MomElim and DotExpand options) from 
  CalcFeynAmp, HelicityME, PolarizationSum functions to
  DeclareProcess.

- Color sums over external particles are now communicated
  to CalcFeynAmp.frm.  They are still not carried out (so as
  not to overcount color when squaring the amplitude) but are
  used to remove terms known to cancel when summed over later.


24 May 16:

- FORM + MathLink binaries are now shipped with FormCalc.
  This should hopefully put to rest the many MathLink build problems.


17 May 16:

- Individual polarization tensors for each leg (eT1, eT2, ...)
  are no longer introduced, rather a function Pol(e, mu, nu) where
  e is the identifier also used for a polarization vector.  That is,
  if 'e' appears alone it is a vector, inside Pol it represents a
  tensor (that made handling in FORM a lot easier).

- The default assumption is that the polarization tensor factorizes
  into two polarization vectors, i.e. eT_{mu nu} = e_mu e_nu, which
  is a specific choice for massless particles, and can be turned off 
  with Clear[Pol].

- Fixed problem with MomElim -> n, which was so far understood as
  a hint to perform dot-product simplification most efficiently
  whereas now it actually eliminates external momentum n.

- Fixed problem with RegisterAbbr which didn't correctly classify
  all external abbreviations.

- MapOnly function added, to apply (e.g.) simplifications only
  to subexpressions containing only particular symbols, for
  example SUSY trig relations.


19 Apr 16:

- Fixed bugs in direct CUT_MREM_... cuts.

- Fixed bug in mktm script for interfacing generated code to
  Mathematica.

- Updated to latest FORM release. 


24 Jan 16:

- Updated detection of Fortran compilers and flags in configure.


22 Jan 16:

- Added PaVeReduce -> LoopTools option which reduces only tensors
  higher than those available in LoopTools.


21 Jan 16:

- Improvements in configure for SIMD detection & pgf* compilers.

- Fixed bug in IndexDelta for negative index values.

- Fixed bug in computation of IDENTICALFACTOR for colored particles.


15 Jan 16:

- Improved detection of libraries (in particular LHAPDF) in configure.

- Added compatibility check for libraries.


13 Jan 16:

- Added UnAbbr function to quickly recover an unabbreviated expression.


11 Jan 16:

- Fixed several problems with vectorization in Fortran/C for different
  values of SIMD.

- Added debug statements to SquaredME.{F,c}.

- Introduced debug level in WriteSquaredME, i.e. no longer on/off
  (#ifdef/#ifndef DEBUG) but graded (#if DEBUG >= 1, 2, 3, 4).


7 Jan 16:

- Added checking statements (DebugLines -> 2 or 3 in PrepareExpr).
  Modified (extended) also the $DebugCmd stuff.

- Added global variables $DebugFF, $DebugAbbr[.], $DebugNum, $DebugRC
  which specify the DebugLines value in the generation respectively of
  the form-factor, abbreviation, numerator, and renconst modules.

- Removed the PARALLEL code from the drivers (util/system/parallel.c).
  Not as efficient as the SIMD code, hence rarely used, and requiring
  a bunch of workarounds for Fortran, too.

- Fixed bottleneck behavior in vplus/vtimes (conversion of C-bound
  expression to vectorized multiplications).

- Updated all examples/*.drivers/run.uuuu.* used for comparison.


24 Dec 15:

- Added computation of effective MU, MD masses, uses DeltaAlfa5Had
  as input.

- Added MB_SCHEME flag to choose OS or MSbar b-mass in model_sm.


22 Dec 15:

- Extended QCD routines so that each routine has its own nloop
  argument.


16 Dec 15:

- Extended {Find,Calc,Write}RenConst functions for arbitrary
  heads, not just RenConst.


11 Dec 15:

- Removed -ff2c flag for gfortran in configure.  Must link with
  LoopTools 2.13 or adjust -ff2c flags in old versions accordingly.


10 Dec 15:

- Added informative printout of helicities being processed, serves
  to check selection of bits in 'helicites' argument.

- Added reset of hseli whenever BIT_RESET is set.


1 Dec 15:

- Added/extended QCD routines.  Formerly in AlphaS.F, now split up
  and moved to util/qcd directory.

- Added cuts.F which implements various cuts by 'veto' (setting the
  cross-section to zero).

- Slimmed down makefiles.


4 Nov 15:

- Added SetLanguage[lang, "novec"] functionality to generate
  unvectorized code only.


3 Nov 15:

- Added new function WeylME and CalcFeynAmp option FermionOrder -> Weyl
  to use matrix-element method also with Weyl chains.

- Fixed many smallish issues with vectorization, mostly in the C output
  (and particularly for clang), mostly when used with HelicityME.


4 Sep 15:

- Added simple way of combining partonic XS, through partonic.h and
  parton.h.


17 May 15:

- New format for squaredme/specs.h allows easier handling of lists
  (e.g. charges in softphoton.F).


4 May 15:

- Updated examples.


28 Apr 15:

- Improvements to ExpandSums and ApplyUnitarity.

- Fixed bug in Chain[VB]7 functions.


1 Apr 15:

- Fixed internals in FermionRC: postponed evaluation of TheMass
  until model is initialized.


31 Mar 15:

- Fixed ToFeynAmp for even simpler expressions.

- Upgraded to latest FORM release.


23 Feb 15:

- Added Chain{V7,B7,V8,B8} to the drivers code.

- Improved tools/USfSimplify.m.


19 Feb 15:

- Introduced functions for DebugLines, e.g. DebugLines -> foo
  individually determines the debug output for each assignment by
  testing whether foo[var -> expr] returns False, True, or a string.


5 Feb 15:

- Added missing value of NVEC, relevant for Cuba integration.

- Increased default value for FCHSELN from 10 to 30 (more conservative).


26 Jan 15:

- Fixed severe error in code generation introduced in Jan 20 release.


20 Jan 15:

- Fixed spurious 0xa0 non-UTF character in FormCalc.m.

- Added SumLegs option for PolarizationSum to restrict the sum to
  a subset of external vector bosons only.

- Fixed compilation issue with the mcc script on Cygwin for
  Mathematica versions before 10.

- Fixed treatment of (rare) 'backslashed' linebreaks in FORM code.


27 Nov 14:

- Upgraded to Cuba 4.1.


4 Nov 14:

- Minor fixes in drivers/tools/fcc for Mathematica 10.0.1.


28 Oct 14:

- Made PolarizationSum honor definitions for eta[i].


19 Oct 14:

- Made $ReadFormHandle publicly available, use it to do
  Uninstall[$ReadFormHandle] before quitting the Kernel to avoid
  (as yet ununderstood) freezes in Mathematica 10.

- Introduced RenConstHook in analogy to InsertFieldsHook and
  CreateFeynAmpHook.


10 Oct 14:

- Fixed technical problem with PaVeReduce which would leave
  denominators containing q1 in the result.

- Split off model-specific definitions to FormCalc/ModelSpecific.m.

- Disabled loading of model-specific defs if $NoModelSpecific = True.


8 Sep 14:

- Fixed comparison of helicities in SquaredME.F, former code
  was correct for unpolarized cross-sections only.


31 Jul 14:

- Unified FV variables in new file model_fv.h, included by
  model_mssm.h and model_fh.h.

- Renamed NMFV preprocessor flag to FV, includes LFV now.
  FV = 2 means sl, su, sd are initialized, FV = 3 only su, sd,
  FV = 4 only sd.


15 Jul 14:

- Various fixes (mostly internal stuff) for Mathematica 10.


10 Jul 14:

- The OffShell function now acts on spinors, too.


8 Jul 14:

- Added OnShell -> ExceptDirac option which avoids using the
  Dirac equation on on-shell momenta.


3 Jul 14:

- Repaired initial build of util.a in compile.


26 Jun 14:

- Fixed problem with shortened routine names in makefile.


4 Jun 14:

- Introduced 'silent' (non-integrating) kinematic variables
  for output only (Show(v) only relevant field), must have
  Step(v) = -999.


2 Jun 14:

- Made build of generated code parallelizable (make -j).
  Still uses libraries, but scripts tools/ar and tools/ranlib
  trick make into building concurrently.


9 Apr 14:

- Added Ninja interfacing code (beta).


7 Mar 14:

- Corrected conjugation properties in ApplyUnitarity.

- Extended OnePassOrder to properly handle IndexIf constructs.

- Extended Folder option of WriteSquaredME to digest composite
  names, such as {"squaredme", ProcName}, where ProcName is
  a (Fortran-, C-, and file-system-safe) name generated from
  the process name.

- Modified makefile.in to handle subdirectory structure underneath
  the squaredme folder, such that several subprocesses will be
  compiled automatically.

- Fixed subtle error in RemoveDups (Optimized -> True in
  PrepareExpr/WriteExpr), where restricted detection but
  unrestricted substitution of subexpressions could lead to
  out-of-the-order "dup" variables.


7 Feb 14:

- Moved to different internal representation for subroutines
  (Nput & friends).  Function is no longer used on the r.h.s.,
  this makes things easier in many ways (due to functions having
  delayed evaluation).

- Straightened out SIMD_* macros, to make the no-SIMD fallback
  version Fortran 77 compliant.


14 Nov 13:

- Added proper initialization with zero of loop result also for
  tree-level-only calculations (relevant mostly for gfortran).

- Reverted to statically linked Linux FORM executables for glibc
  compatibility.  Static linking had caused segfaults in g++'s
  exception handling (used in FORM's factorization algorithm).
  In the event of a segfault in FORM please replace
  Linux{,-x86-64}/tform by the respective shared version
  FORM/Linux{,-x86-64}/bin/tform-shared.

- Introduced FCHSELN and FCHSELEPS environment variables to detect
  and eliminate negligible helicity combinations.


13 Nov 13:

- Updated FORM executables.

- Added Windows-x86-64 build (on Cygwin64).

- Made detection of helicity combinations with zero result more
  conservative.


8 Nov 13:

- NCOMP = 1 now automatically sets BIT_LOOP = 0.


7 Nov 13:

- Fixed glitch in WriteSquaredME for vanishing tree (or loop)
  contribution.


6 Nov 13:

- configure now creates simd.h (if different than before) such
  that dependencies are properly handled.  This affects mostly
  util/kin/VecCopy.F.


3 Nov 13:

- Fixed MathLink compilation issue with MacOS 10.9/clang++.

- Enabled vectorization as default using the capabilities of the
  CPU the code is configured on.  Use ./configure --generic to
  turn off, i.e. make code run regardless of CPU features (e.g.
  for batch processing on a cluster).

- Added automatic detection of helicity combinations with
  zero result.


9 Sep 13:

- Fixed detection of LHC energies in lumi_hadron.F.


8 Sep 13:

- Fixed another problem with categorization of abbreviations
  (MoveDepsRight this time).


5 Sep 13:

- Upgraded to Cuba 3.2.  Compilation with clang possible now.


1 Sep 13:

- Fixed problem with categorization of abbreviations (MoveDepsLeft).


2 Jul 13:

- Fixed generation of specs.h for particles with missing quantum numbers.

- Fixed minor issues in 2to4.F.


23 May 13:

- Improved simplification of OPP numerators.


8 May 13:

- Fixed technical problems in compiling OPP code with SIMD > 0.
  Current versions of Samurai or CutTools allow only SIMD = 1,
  however.

- Fixed some compiler issues in C code related to modification of
  spinors of May 6.


7 May 13:

- Fixed FORM hiccup with Truncated -> True.


6 May 13:

- Fixed problem with spinors in the presence of self-conjugate fermions.

- Upgraded to Cuba 3.1.


23 Apr 13:

- Vectorization for helicity loop added in Fortran and C, in C uses
  SSE3 (for SIMD = 1) or AVX (for SIMD = 2).  Due to vectorization,
  the fixed-format Fortran output is effectively Fortran 90 now. 
  Fortran 77 can still be forced if SIMD = 0 in distrib.h.
  Thanks to Jean-Nicolas Lang for the SIMD implementation in C and
  extensive testing.

- For the moment, Mac users need to ./configure CC=clang for SIMD > 0,
  due to the availability of immintrin.h.  Caveat: the Cuba part of
  util.a does not (yet) compile with clang because of clang restrictions
  on C99 language features.

- Parallelization (#define PARALLEL 1 in distrib.h) extended to C.

- New file distrib.h controls vectorization and parallelization for
  both Fortran and C.

- The properties of the external particles, which so far had to be
  hand-coded in process.h, are now generated (squaredme/specs.h).
  Old process.h files have to be rewritten to include this file
  instead of the direct definitions.


27 Mar 13:

- Fixed a bottleneck in CalcRenConst.

- Added FileTag option to CalcFeynAmp.


25 Mar 13:

- Fixed a few formatting statements for Mathematica 9.

- Added new onlyfunctions option to FORM code.


11 Mar 13:

- Added drivers for C code, restructured 'drivers' directory
  (C and F subdirectories) and makefiles.


22 Feb 13:

- Fixed bugs in the formal arguments of the 'Fewest' procedure
  in HelicityME.frm and PolarizationSum.frm introduced in the
  last release of Feb 7.

- Introduced wrappers RCSub, RCInt for simplification of the
  renormalization constants (analogous to FormSub etc.).

- Improved variable write-out in WriteSquaredME.  Added NameMap
  output mode (similar to Common).

- Improved numerator simplification for OPP.

- Fixed two memory-allocation bugs in ReadForm which could
  occur for large amplitudes.


6 Feb 13:

- Fixed problem with unquoted commas in preprocessor
  statement in invariant simplification.


5 Feb 13:

- Fixed PolarizationSum for GaugeTerms -> False:
  momentum conservation must be applied *before* dropping
  gauge-dependent terms or else not all cancellations are
  taken into account.


1 Feb 13:

- More optimizations on OPP denominator extension.

- Fixed subtle bug that would lead to dot products
  not being abbreviated and ultimately wrong results.

- Overhauled OnSize function.

- Fixed problem of neglecting masses in loop integral arguments.


21 Dec 12:

- Fixed bug in parallel Divonne (ncomp > 1).

- Removed Source option in HelicityME, ColourME, and PolarizationSum.
  Too awkward syntax with both Abbr[] and Subexpr[].

- Putting the OPP integrals on a common denominator is now done in
  Mathematica, this is both easier and works across parts of the amplitude.
  One should start with the largest N first, e.g. the three-point OPP
  functions will be extended to four-point ones only if the vertices are
  computed after the boxes.


12 Dec 12:

- Split Abbreviate function into AbbrevSet and AbbrevDo, can introduce
  subexpressions as they come from FORM with this.

- Fine-tuned simplification procedures, there are now:
  FormPre to set up things (if necessary),
  FormSub for the subexpressions introduced by FORM (subM),
  FormDot for expressions involving dot products (dotM),
  FormMat for the coefficients of the Mat[...] matrix elements,
  FormNum for the numerators in OPP.

- Updated HelicityME and PolarizationSum with the FORM 4 features
  already used in CalcFeynAmp.


22 Nov 12:

- Rewrote procedure DotSimplify in CalcFeynAmp.frm, better selection
  of terms for simplification hence faster.

- Moved to tform with $ProcessorCount workers.  Not (yet) subtracting
  load average.


6 Nov 12:

- Improvements to the abbreviationing system.

- Introduced FormSubSimplify and FormMatSimplify for finer-grained
  simplification in Mathematica.

- New TreeCoupling function.


24 Oct 12:

- Moved to FORM's ToPolynomial abbreviationing system.

- Modified ReadForm quite a bit to account for abbreviations.
  Removed MAXEXPR hassle en passant.

- Optimized the position of some .sorts in CalcFeynAmp.frm.


6 Oct 12:

- Rearranged internal handling of loop integrals, FORM gives back
  paveM and cutM, N-specific renamings happen in Mathematica.

- The cut integrals now carry a set of 'traditional' parameters
  (with invariants), as in: Ccut[c0para][ccutpara], for the
  purpose of separating off calculation of the master integrals
  in the generated code (to be hoisted out of the helicity loop).

- Fixed problems with 'if' and 'call' syntax in C.


20 Sep 12:

- Fixed dependence problem in *put abbreviations.

- Added latest FORM executables.


18 Sep 12:

- Changed CancelQ2 to SimplifyQ2 which includes merging lower-N
  with higher-N OPP integrals, as in:
  N2/(D0 D1) + N3/(D0 D1 D2) -> (N2 D2 + N3)/(D0 D1 D2)


12 Sep 12:

- CalcFeynAmp hands down full process info now, not just masses.

- Fixed problem with DiagramType for Propagators appearing in
  higher powers.


8 Sep 12:

- Helicity loop runs parallel now if PARALLEL is defined,
  e.g. in user.h.  Currently takes all free cores (all cores 
  minus system load), this can be changed by setting the
  environment variable FCCORES.  The number of cores Cuba takes
  is not (yet) taken into account.


22 Jul 12:

- Improved C output.  Using C99 declarations for initializing
  the Mat... with NaNs.

- Split helicity loop of SquaredME into own routine SquaredMEHel.

- Made all 'util' functions inline (inline.h), for better performance.

- Improved VarDecl function, now accepts strings which are passed
  verbatim (such as "#ifdef ...", "#endif").  Conventions for
  common blocks have changed: used Common[name][vars].


25 May 12:

- Added Cuba 3.

- Added MomRules option to CalcFeynAmp.


12 May 12:

- Fixed problem with (4D) Fierz identity: now simplification of
  fermion chains is invoked after application of each Fierz id,
  not once at the end.

- Added CalcFeynAmp option Gamma5Test.

- Added CalcFeynAmp option Gamma5ToEps.

- Added CalcFeynAmp option OPPQSlash.

- Added CalcFeynAmp option Evanescent which keeps objects
  of the form Evanescent[op before Fierz, op after Fierz] in
  the amplitude, e.g. to track down D - 4 issues.


16 Mar 12:

- Added macros for better abstraction of loops and if statements.

- Added PowerOf function which aids splitting of amplitude into
  pieces of the same power in the coupling constant.

- Replaced stacked assembly of elementary operations in the
  computation of spinor chains by single function Chain.


12 Jan 12:

- Speeded up analytic tensor reduction (PaVeReduce -> True).


23 Dec 11:

- Fixed problems with index-dependent numerators.

- Fixed 64-bit issue (ptrdiff_t casted to int) in ReadForm.tm.


2 Dec 11:

- Moved from Real|Complex to RealType|ComplexType since the former
  is understood by the compiler and does not lead to warnings.

- Fixed bug in index dependence of numerator subroutines and
  derived abbreviations.


25 Aug 11:

- Introduced types 'Real' and 'Complex' (defined in types.h) for
  better abstraction in Fortran.

- Split common blocks in vars.h, e.g. abbrev splits into
  abbrev0s, abbrev0angle, abbrev0hel etc.


16 Aug 11:

- Corrected wrong index distribution in contraction in SUNEps.
  (Thanks to Sho Iwamoto.)


8 Aug 11:

- Modified the OPP option of CalcFeynAmp: now accepts integer (former
  values True, False, Rational in fact mapped to integers now) indicating
  from which N on an N-point function shall be treated with OPP methods.


4 Aug 11:

- Updated the CutTools drivers to match invocations in the latest version
  of CutTools.

- Updated the diagonalization routines to the latest version.


20 Jul 11:

- Fixed problem with escaped backslashes in ReadForm, as in \\[Eta].

- Added CalcFeynAmp option CancelQ2 to be able to switch off cancellation
  of q^2-terms in the numerators.


11 Jul 11:

- Refitted mktm with argc and argv arguments.

- mktm now captures stdout and sends it to Mathematica.


22 Jun 11:

- Added tags (comments) to SquaredME.F for easier sed editing.


10 Jun 11:

- Fixed handling of Mathematica's List in FORM.


30 May 11:

- Added PreFunction and PostFunction options to CalcFeynAmp.

- Added TagCollect function.


27 May 11:

- Added SortDen option to CalcFeynAmp, this is currently still
  needed for Samurai but may generally be useful for testing.


12 May 11:

- Fixed sign in PaVeReduce code.  (Thanks to A. Neiser.)


11 May 11:

- Fixed PolarizationSum bug for longitudinal reference vectors
  in fermionic amplitudes (the s[n] did not get converted to
  e[n] when passing to FORM).


5 May 11:

- Fixed absorption of MuTildeSq in the abbreviations for OPP -> True.

- Updated constants in model_sm.


7 Apr 11:

- Made PrepareExpr aware of nested instances of FortranExpr.


6 Apr 11:

- Fixed problem with open Lorentz indices in conjugate eps tensors.

- Fixed problem with sum in FORM code, replaced all dummy indices +
  sum by N100_? and renumber.


5 Apr 11:

- Fixed quotes around `m' in PolarizationSum.frm, re-enabling
  detection of massless legs.

- Added optional 'colon arguments' (:arg1 :arg2) to main.F, these are 
  passed to ModelDefaults and LumiDefaults.


1 Apr 11:

- Made LoopIntegral, PaVeIntegral, CutIntegral public.


24 Mar 11:

- MkDir returns path ending in /, for better compliance with
  $PaintSE etc.


14 Mar 11:

- Fixed memory allocation problem in ReadForm.


11 Mar 11:

- Fixed summation range in PolarizationSum.frm for arbitrary
  dimensions.


7 Mar 11:

- Renamed integration variables for better consistency across
  phase-spaces with different number of legs.

- Fixed invocation of Cuba (segfault).

- Fixed C_{mu nu} local term in DRED.

- Added EditCode -> Modal to allow modification of FORM code.


10 Feb 11:

- Replaced calls to Mathematica's Unique by NewSymbol, with its
  own numbering system.  Unique's numbering is just too
  unpredictable in Mathematica 8.


8 Feb 11:

- Cleaned up the NMFV sections in model_mssm.F.
  Due to copy-and-paste from FeynHiggs, some bits were not compiling.


5 Feb 11:

- Fixed FORM error with NEQ (in DRED local terms).

- Extended cancellation of q1.q1 to handle several instances.

- Reverted to plain numbering of dups and tmps in PrepareExpr
  (rather than using Unique).  In repeated calls not so many
  different symbols are generated thus.  Use DupOffset and
  TmpOffset to obtain numbering not starting from 1.


3 Feb 11:

- Added -ff2c to gfortran FFLAGS for compatibility with other libs.


2 Feb 11:

- Corrected powers in reduction of A00 etc.

- Fixed infinite loop if several instances of q1.q1 were around
  (does not affect regular one-loop calculations).


1 Feb 11:

- Fixed problems with pattern [m1] in one-point tensor reduction.

- Added renumber statement to make FORM try harder to find
  identical terms under sums.


31 Jan 11:

- Fixed local terms in DRED.

- Made configure find lib64 directories (Linux x86-64).


21 Jan 11:

- Fixed problem in ReadForm.tm which would lead to loop integrals
  being dropped from PolarizationSum output.

- Switched typedefs for Bget, Cget, Dget, Eget from integer to
  memindex, in accordance with latest change in LoopTools for
  64-bit compliance.


16 Dec 10:

- Again lots of rearrangements in the numerator functions.
  Now has macros 'NumeratorFunction' and 'Result'.
  Spinor chains get evaluated on one line each, to prevent
  the Fortran compiler from evaluating them in the wrong order.

- Moved Abbreviate[expr, patt] to Abbreviate[expr, foo] (like Select).

- Subexpr[args] executes Abbreviate[args] locally, i.e. without
  registering the abbreviations permanently.

- Added Fuse option of Abbreviate.  Fuse -> False introduces
  abbreviations for elementary items only.


6 Dec 10:

- Lots of rearrangement re CutTools/Samurai.

- Added util/samurai directory with wrapper functions.

- decl.h and user.h are now both file includes and subroutine
  includes, i.e. the first #include just declares the per-file
  things, like preprocessor #defines, and only on subsequent
  #includes are Fortran declarations emitted.

- Renamed CutTools option of CalcFeynAmp to OPP.


1 Dec 10:

- num.h is now a link to either cuttools.h or samurai.h, both of
  which need the numerator in a slightly different form.


2 Oct 10:

- Removed model.h in favour of model_{sm,mssm,thdm,...}.h, which
  allows to have separate declarations for every model.  The new
  model_xxx.h should be included in user.h.


23 Sep 10:

- Added Dminus4Eps for the Dminus4 coming from epsilon contractions
  in HelicityME and PolarizationSum.  This is not very consistent.
  Mostly for testing, I surmise.  You have been warned.

- Moved External.h into CalcFeynAmp.frm, as now the proper
  Cygwin FORM executable is available.  External channels now
  available on Cygwin!


21 Sep 10:

- Windows FORM executable now a properly compiled Cygwin one.


16 Sep 10:

- Fixed remaining source of g_(1) (in fill statements) in
  HelicityME.frm, would lead to grossly wrong results.


15 Sep 10:

- PaVeReduce -> True, Raw working (beta).  Integrals with N > 4
  not checked yet.


10 Sep 10:

- Fermion chains in HelicityME are now written out before
  taking the traces.  Could lead to wrong D-4 terms in the
  presence of two epsilon tensors.


9 Sep 10:

- Fixed issue with lack of gamma matrices in fermionic couplings,
  e.g. fermion chains of the form Spinor(...)*Spinor(...).


3 Sep 10:

- Added Dimension -> 0 for HelicityME and PolarizationSum, too.


11 Aug 10:

- Fixed problem with the abbreviation of negative powers of
  dot products: negative powers must be included in the abbreviation 
  on the FORM side and Mathematica pulls then out again.
  Exact reason for wrong results remains unclear.  Problem affects 
  PolarizationSum only, as the other routines do not deal with
  negative powers.  (Thanks to Celine Degrand.)

- Fixed issue with including static FORM code before
  generated code, as in HelicityME.frm, PolarizationSum.frm
  (uses ## delimiter also after static code).


5 Aug 10:

- Consistently moved all strings in ReadForm.tm to unsigned char,
  using ML{Get,Put}ByteString for transfer.

- FormCalc now includes the actual FORM code in the output file,
  i.e. resolves the #include.


4 Aug 10:

- Fixed up a lot of Windows anomalies, e.g. properly quoted
  filenames in the sense of Cmd.exe.


23 Jul 10:

- Added SUSYTrigExpand, SUSYTrigReduce, SUSYTrigSimplify
  for simplification of the various symbols representing
  SUSY angles, such as S2A, TB2, ...

- Added MassDim function to check for the correction mass
  dimension of an expression.


20 Jul 10:

- ToDoLoops now takes care to order the expressions such
  that e.g. a loop that sums up variables comes after the
  loop calculating them.  (Thanks to Martin Wiebusch.)

- Added FinalCollect option to WriteExpr, was executed
  automatically so far but could take quite long.

- OpenFortran passes options to OpenWrite.


14 Jul 10:

- Fixed a lot of hiccups on Ubuntu.  Disabled optimization for
  gcc 4.2.x and 4.4.3, which would lead to an Internal Compiler
  Error when compiling Cuba.


30 Jun 10:

- Fixed hideous bug in ToFortran.c (strcpy vs. memmove) which
  would lead to malformed Fortran output on some platforms.
  (Thanks to Alex Arbey.)


17 Jun 10:

- Fixed problem with ClearProcess: the zap function (Remove) was
  not applied correctly, so numbering of abbreviations and 
  subexpressions did not get reset.

- Repaired problems with the substitution of vectors in
  ToComponents, nested spinor products, and conjugation properties
  of polarization vectors in VecSet.m.


11 May 10:

- Improved GaugeTerms -> False to include originally eta-dependent
  terms where the eta dependence drops out due to cancellations.

- Fixed 'mysterious' problem that abbreviations would be
  generated twice, rather delicate handling of head vs. whole
  function substitution in SplitExpr.


4 May 10:

- Kept IndexSum from evaluating inside the G[_][__][_][_]
  expressions, renaming index would upset substitution of
  insertions.

- Made Lorentz index in FourVector optional, to be used as in
  LeviCivita[mu, nu, FourVector[k], FourVector[p]].


20 Apr 10:

- Added ExtractInt to isolate invocations of LoopTools functions
  from an expression, i.e. the B0i, C0i, ... invocations are
  replaced by Bval, Cval, ... and a list of corresponding
  Bget, Cget, ... calls is added.


15 Apr 10:

- Fixed various issues with explicit Dirac indices.  Ghosts
  either carry DiracObject[1][...] with them or P$NonCommuting
  has to include U.

- Added IndexSum[..., x] (not {x, n}) to indicate summation
  over Lorentz indices.


26 Mar 10:

- Explicit Dirac indices on fermion objects are now handled,
  in-chain: DiracObject[DiracMatrix[...]][di1, di2],
  end-of-chain: DiracObject[DiracSpinor[...]][di],
  See Dirac.gen of FeynArts.


26 Feb 10:

- Fixed problem with transferring Mathematica's special symbols
  like \[Alpha] to/from ReadForm.


3 Feb 10:

- Fixed problem with (differences of) vectors becoming zero
  in arguments of loop functions.


31 Jan 10:

- Added .so files to the linker-relevant bits in fcc.


16 Dec 09:

- Fixed problem with universal binaries (Mac OS) in fcc.

- Added MmaGetInteger, MmaGetIntegerList to mktm.


27 Nov 09:

- Added {SM,MSSM}{Reduce,Shorten,Simplify} functions for
  model-specific simplification.


13 Nov 09:

- Re-introduced "Divergence" in UVDivergentPart.  The version
  with Dminus4, being in the denominator, was performing 
  significantly worse in the simplification of renormalization
  constants.


10 Nov 09:

- Added NClear function to remove numerical values for constants.


4 Nov 09:

- Added NoBracket option for CalcFeynAmp.


12 Oct 09:

- Removed hideous bug from bit encoding of helicities for
  external massive vector bosons.

- Added Dimension -> 0 mode where the dimensionality is retained
  (Dminus4 appears in the amplitude).  Also no local (rational)
  terms are added.


2 Oct 09:

- Fixed problem with Generic-only insertions.

- Avoided substitution of momenta in CalcFeynAmp if those were
  not of type FourMomentum.  This gave rise to silly errors
  when people set momenta to zero by hand.


3 Sep 09:

- Added |Hel| = 2 modes for graviton and gravitino.  Not complete
  yet, i.e. wave functions not initialized.  Just the "helicities"
  argument of SquaredME.


1 Sep 09:

- Fixed problem with conjugation of VA chains in HelicityME.


27 Aug 09:

- Moved sorting of propagators inside loop back to FORM,
  otherwise InsertionPolicy -> Begin would never have any effect.


12 Aug 09:

- Bug fixes in SubstSimpleAbbr and inv.

- DiracCoeff glosses over polarization vectors
  (experimental, for spin-3/2 particles).


30 Jul 09:

- Temporary FORM file is written out with explicit path now.
  Was a problem when doing a SetDirectory after loading
  ReadForm (FormCalc).


29 Jul 09:

- Fixed linking problem on Mac due to gcc -lstdc++ being used
  instead of g++.


27 Jul 09:

- Yet more changes re Windows.
  Fixed some bizarre problems in ReadForm for Cygwin with strings
  and fork/exec.


22 Jul 09:

- Lots of trivia re Windows version, such as:
  quoting for filenames with spaces,
  removal of $TemporaryPrefix (%TEMP% not readable for non-root),
  deactivation of -pipe switch for Win FORM.


6 Jul 09:

- Fixed problem due to overzealous replace_ in DotSimplify
  (would work on e.g. Spinors).


30 Jun 09:

- Changed MomSimplify to MomElim: user can now choose which
  momentum to eliminate via momentum conservation.  Automatic
  chooses the combination that generates the fewest terms.


22 Jun 09:

- Replaced "contract 0" in CalcFeynAmp.frm by a procedure which
  contracts only the e_ connected through Lorentz indices (thus
  generating fewer terms).


19 Jun 09:

- Revamped DotSimplify in CalcFeynAmp.frm to use the shortest
  combination of vectors for each term individually.


16 Jun 09:

- Adapted conventions of VecSet.m to VecSet.F and fixed a bug in 
  the application of spinors in VecSet.m (missing multiplication 
  with Hel[i]).


15 Jun 09:

- Fixed problem with e_ and composite vectors (e_(p1+p2, ...)).

- Added proper contract statements in DotSimplify.


12 Jun 09:

- Added the possibility to pass options to InsertFields to
  SelfEnergy, DSelfEnergy, MassRC, FieldRC, TadpoleRC, WidthRC.


9 Jun 09:

- Cleaned up some context associations for compatibility with
  the LoopTools MathLink executable.

- Fixed cancellation of factors involving gauge parameters after
  partial fractioning of denominators.


3 Jun 09:

- Fixed minor bits in FORM procedure "Square" for Mac FORM.


28 May 09:

- Taken into account `hidden' index dependence through SumOvers and
  lower-level DoLoops in Hoist ("Expensive" option of WriteExpr).


26 May 09:

- Fixed quoting problem with explicit Fierz ordering in
  CalcFeynAmp.frm.


23 Apr 09:

- Added the RealArgs *function* for fine-tuning the conversion of
  integers into reals for WriteExpr.  Cases in point: Acut, Bcut, etc.


22 Apr 09:

- Added the Expensive option to PrepareExpr.


1 Apr 09:

- Added Mma{Get,Put}{Real,Complex}List functions to mktm.


26 Mar 09:

- (Hopefully) fixed the spurious detection of 32-bit systems
  as 64-bit ones in configure.


25 Mar 09:

- Replaced MLReleaseString by MLDisownString in ReadForm.tm,
  the former being included in the MathLink of Mathematica
  versions 6 and up only.


20 Mar 09:

- Added the Sq[...] =. definition.


15 Mar 09:

- Added the $LongitudinalSE variable with which the longitudinal
  part of the self-energy can be selected.


12 Mar 09:

- Extended UVDivergentPart to recursive definitions of RCs
  (this holds for the currently defined RenConst).


11 Mar 09:

- Added $LongitudinalSE for taking the longitudinal part of the
  self-energies.


5 Mar 09:

- Re-enabled pipe communication for Mac OS and fixed up some
  details in the handshake.  Works for most cases now.

- Patched a couple of warnings in ReadForm.tm.


23 Feb 09:

- Fixed problem with negative powers of functions inside
  fermion chains in FORM.


20 Feb 09:

- Added ZapFunction option to preserve symbols across
  ClearProcess.

- Added workaround for MacOS FORM (disabled pipe communication).


11 Feb 09:

- Fixed bug in categorization of abbreviations (problem for
  spin vector s(i) only).


28 Jan 09:

- Corrected signs in charge conjugation of Dirac chains during
  Fierzing, must be same as in HelicityME.

- Added nicer formatting for Dirac and Weyl chains.


22 Jan 09:

- Fixed problem in HelicityME: negative chiralities were
  flipped wrongly in ConjChain (e.g. -7 xor 1 = -8, not -6).

- Lorentz indices arrived in Mathematica as Lor[2$] instead
  of Lor[2] due to a mistake in the character handling.


13 Jan 09:

- Fixed compile and fcc for Mathematica 7.


7 Jan 09:

- DiracChain[om, mu, nu, ...] -> DiracChain[-om, mu, nu, ...]
  if Antisymmetrize -> True.


2 Jan 09:

- Added FileIncludes, SubroutineIncludes, FilePrefix, and
  FileHeader options to both WriteSquaredME and WriteRenConst.


20 Dec 08:

- Again changed _ substitute to $$ (not ß), because of problems
  w/character encoding on Mac.

- Added Abbreviate[expr, patt] functionality.


17 Dec 08:

- Internal "escape character" for _ in FORM output is now ß
  (_ + 128).  This avoids clashes with Mathematica's auto-generated
  patterns, e.g. i_ becomes i$_ if definition is created from inside
  a function.

- Replaced the one remaining hel by helM in HelicityME.frm.


16 Dec 08:

- Fixed (trivial) bug in VecSet.m.


15 Dec 08:

- Introduced ...M notation for symbols shared between FORM
  and Mathematica to minimize symbol collisions.


12 Dec 08:

- Fixed race condition in threads handling pipe communication.

- Added Antisymmetrize option of DeclareProcess.


10 Dec 08:

- Added the TreeSquare option of WriteSquaredME.

- Fixed two bugs in the pipe code of ReadForm.


3 Dec 08:

- Hopefully final correction of signs in fermion-flip rule in
  HelicityME.frm.

- Moved test of BIT_RESET in SetEnergy such that ModelVarIni is
  not called every time.

- Reformulated AbbrScale-related code without $-variables.

- Added fortranflush function to util/system/ and mktm for
  compatibility with gfortran.


27 Nov 08:

- Replaced popen call with explicit pipe & fork.  Now sending an
  explicit kill to the FORM process before wait, thus reliably
  handling aborts.


25 Nov 08:

- Fixed problem with Majorana spinors and HelicityME.
  Also e_ were wrongly abbreviated inside fme.


20 Nov 08:

- Extended IndexIf for the "else if" cases, i.e.
  IndexIf[c1, a1, c2, a2, b] corresponds to
	if( c1 ) then
	  a1
	else if( c2 ) then
	  a2
	else
	  b
	endif


17 Nov 08:

- Added the Keep function.


14 Nov 08:

- Renamed InsertAt to InsertionPolicy and added integer values,
  specifying a LeafCount above which the insertion is postponed
  until the amplitude comes back from FORM.

- Fixed problem with FindRenConst: expansion of RCs has to
  happen during the search process, not at the end.


13 Nov 08:

- Moved filling-in of insertions completely to replace_.


11 Nov 08:

- Added explicit variable declaration to pipe I/O, to
  safeguard for user definitions and other hiccups.


31 Oct 08:

- Fixed bug in RegisterAbbr (ferm set without scale argument).


30 Oct 08:

- Made DeclareProcess an independent and public function.
  Useful e.g. if one wants to compute HelicityME without
  going through CalcFeynAmp.


29 Oct 08:

- Accounted for the antisymmetrization also in HelicityME.
  Moved much of the HelicityME code to FORM (HelicityME.frm)
  in the process.  Gained some efficiency on bracketing out
  helicities cleverly for massless projectors.


15 Oct 08:

- Added alternate-channel communication in ReadForm.
  Amplitudes are sent to Mathematica for abbreviationing
  (usually) after generic simplification.

- DiracChains are now antisymmetrized, i.e. DiracChain[1, mu, nu]
  is Sigma[mu, nu].


31 Jul 08:

- Introduced Error(...) and Warning(...) together with Error.F
  in a FeynHiggs-like error-reporting system, i.e. suppresses errors
  repeating fast.

- Separated subroutine HiggsMasses in model_mssm.F into
  HiggsMassesConst and HiggsMassesVar.


30 Jul 08:

- Eliminated a few bugs from the NMFV initialization in model_mssm.F.

- Added RegisterAbbr function similar to RegisterSubexpr.


29 May 08:

- Added FORCE_ONSHELL to lumi_parton.F, to force a decaying
  particle on-shell no matter which sqrtS is given from the
  outside.


13 May 08:

- Modified PrepareExpr to deliver a "fully Mathematica"
  output in FortranExpr.  DoLoop and IndexIf structures are
  finally translated into Fortran code only by WriteExpr.

- Added DebugLines option to PrepareExpr and corresponding
  function DebugLine.


8 May 08:

- Improved CalcRenConst: no indiscriminate collection
  w.r.t. integrals any more, only when necessary (gives
  shorter expressions).


30 Apr 08:

- Added Abbr[patt] (i.e. with pattern).

- Added Deny option of Abbreviate.


25 Apr 08:

- Complete overhaul of the WriteExpr function.  Rolled
  WriteDoLoops and WriteSummedExpr into WriteExpr.


20 Mar 08:

- Extended the ColourFactor function to ColourSimplify,
  for public use.

- Added the ColourGrouping function, which serves as an
  extension to FeynArts' DiagramGrouping function in that
  it simplifies colour structures before grouping.


2 Mar 08:

- Introduced IndexDiff[i, j] = 1 - IndexDelta[i, j].

- Added ToIndexIf function to turn IndexDeltas and
  IndexDiffs into IndexIfs.

- Modified WriteExpr and WriteSummedExpr to correctly
  generate code containing IndexIfs which are translated
  into Fortran if-statements.


28 Feb 08:

- Added LeviCivita symbol for antisymmetric tensors with
  kinematic indices appearing in the generic model file.


26 Feb 08:

- Fixed a couple of quirks in main.F (among others) mainly
  for the Alpha platform.

- Added setfpu.c to put the FPU into trapping mode (i.e.
  floating-point errors are immediately fatal) for g77 and
  gfortran.


20 Feb 08:

- Fixed a bug in ApplyUnitarity.


11 Feb 08:

- Removed DECL1..20, unnecessary now due to user.h.


4 Feb 08:

- Added generation of if statements (IndexIf) - useful e.g. for
  RCs dependent on index equality (dZf etc.).


18 Jan 08:

- Added decl.h as one-stop include file for all generated code.

- Added separate file user.h for user declarations.


27 Nov 07:

- Improved compile script.

- Fixed SetEnergy subroutine to call ModelVarIni even if the
  energy is the same but the RESET flag is set.


23 Nov 07:

- Fixed xlf-problem in Diag library (wouldn't accept 2D0**-102,
  must be 2D0**(-102)).


22 Nov 07:

- Added FORM binaries to distribution.  Fed up with advising folks
  to use latest version.


20 Nov 07:

- 'Cleaned up' some internals: new function KinFunc which
  allows to cast statements like #3.#4 -> -1 as e.ec -> -1
  (mainly in DeclareProcess).

- Enlarged the FORM setup parameters, in particular
  MaxTermSize.  This is irrelevant for 32-bit platforms,
  where the maximum is 32568, but useful for 64-bit ones.

- Downgraded a few makefile.in features for compatibility
  with GNU make 3.80.


18 Nov 07:

- Fixed bug in the normalization of external polarization vectors.
  (Was ei.eci -> 0, not -1, due to copy & paste error.)
  Relevant only for HelicityME.

- Improved PolarizationSum by introducing the gauge-dependent
  vectors eta explicitly (formerly e(i)).  Added calls to eiei
  and eiki routines, should speed up slightly.


11 Nov 07:

- Extended CalcFeynAmp's InsertBefore flag to InsertAt, with choices:
  1. Begin = apply Insertions in the FORM code, but before
     simplifications,
  2. Default = apply Insertions in the FORM code, but after
     simplifications,
  3. End = apply `bulky' Insertions in Mathematica
     (workaround for FORM bugs).


8 Nov 07:

- Added DECL1..20 statements to run.F for declarations of local
  variables.

- Made DECL*, LOOP*, PRINT* statements optional (i.e. the user
  may omit superfluous ones).

- Fixed the drivers/tools/mcc script for Mathematica 6 under
  Windows/Cygwin.

- Upgraded to latest version of the Diag library (1.3).


7 Nov 07:

- Further simplified generated makefiles.  Macros DIR and LIB
  of makefile.in are now used.

- Added SqrtPhiGl definitions for the gluino phase.


18 Oct 07:

- Fixed bug in ColourME: for powers of a colour trace, e.g.
  SUNT[Glu2, Glu1, Glu3, 0, 0]^2, only one replacement index for
  the "0, 0" was generated.  This led to a wrong colour factor.


17 Oct 07:

- Moved flush calls from logfile.c to xsection.F, due to special
  handling of the flush "intrinsic" by gfortran.


11 Oct 07:

- Upgraded the FeynHiggs interface to the latest version (2.6.1).


15 Sep 07:

- "Downgraded" to 32-bit compile on x86_64 systems for compatibility.


12 Sep 07:

- Added charge conjugation to the FermionOrder techniques.
  Fermion chains are now definitely rearranged in the prescribed
  order.


29 Aug 07:

- Replaced USfSimplify by the more general function
  ApplyUnitarity.

- Since USfSimplify might be slightly more thorough, it is
  still available in tools/USfSimplify.m.


28 Aug 07:

- Improved 4D Fierz identitites (better matching, more cases).

- Introduced Sigma[mu, nu] on 4D fermion chains.


14 Aug 07:

- Renamed HornerForm (option of CalcFeynAmp) into HornerStyle.
  FormCalc is now fully compliant with Mathematica 6.


9 Aug 07:

- Modified the compile and mcc script for compliance with Mma 6.


2 Aug 07:

- Added EpsilonScalar indices and corresponding polarization
  vectors z, zc.

- Renamed the polarization tensors from ee, eec to eT, eTc because
  of collisions with the ee* symbols of the 5-point function.


20 Jul 07:

- Again changed makefile setup for compliance with older gmake
  versions (< 3.80).


18 Jul 07:

- Changed SPEC_K{X,Y,Z} to SPEC_E{X,Y,Z}.  The SPEC_Ki can
  easily be reconstructed from SPEC_K and SPEC_Ei.


13 Jul 07:

- Renamed Small to Neglect for compatibility with Mma 6.


11 Jul 07:

- Revamped makefile handling in the drivers to avoid absolute
  paths.  These are a nightmare for make if they contain spaces,
  which is often the case on Windows and Mac OS.

- Made compilation process Windows compliant.  FormCalc would
  now run under Windows (Cygwin) if a FORM version for Cygwin
  were available.


1 Jul 07:

- Added handling of polarization tensors for gravity in the
  symbolic part.


8 Mar 07:

- Added Fierz identities for 4D spinors.  The order in which
  the external spinors shall be arranged is chosen with the
  FermionOrder option of CalcFeynAmp.


26 Feb 07:

- Added SEPart function to allow for modifications of parts of
  renormalization constants.


13 Feb 07:

- Added MassRC, FieldRC, TadpoleRC, WidthRC functions to simplify
  the definition of renormalization constants (and introduce
  fewer errors).


6 Feb 07:

- Added MoveDepsLeft and MoveDepsRight functions to disentangle
  dependences between categories of abbreviations.


3 Feb 07:

- Separated abbreviations into such for tree and loop part.
  Changed the third "reset" argument of SquaredME to a more
  general flag register.  Bit 0 is the old "reset" value,
  bit 1 determines whether the loop correction should be added,
  i.e. bit 1 = 0 computes the tree-level part only.


29 Jan 07:

- Added ToArray function which turns symbols into array references,
  e.g. Xyz1234 -> Xyz[1234].

- Added MaxDims function which determines the maximum dimensions of
  array references, e.g. MaxDims[Xyz[1, 2], Xyz[2, 1]] -> {Xyz[2, 2]}.

- Added Renumber function which renumbers the given array references.

- Added SubstSimpleAbbr function which removes `simple' abbreviations
  and subexpressions from the list and substitutes them directly
  into the expressions they appear in.


24 Jan 07:

- Added startmath template script with which programs that use the
  Mathematica interface can be distributed (parallelized).

- Added handling of IndexEps objects which represent antisymmetric
  combinations of indices (e.g. colour or generation).


15 Jan 07:

- Revamped ReadForm.tm for better C style.  Adapted the MathLink
  string functions to the latest MathLink version.

- Fixed configure to work properly with pgf77 6.2.

- Fixed some problems with the Mathematica interface.


18 Nov 06:

- Added Mac OS X details to configure and makefile.in for g77 and xlf.


30 Oct 06:

- Added the six-point tensor reduction.

- Fixed overflow problem in FORM due to the procedure DotSimplify
  in CalcFeynAmp.


24 Oct 06:

- Added Mathematica interface for generated code.
  Generalized some sections of CalcFeynAmp.frm, added
  the F function as far as tensor reduction goes.


11 Sep 06:

- Adapted FeynHiggs interface to latest version (2.4.x).

- Adapted LHAPDF invocations to latest version (5.x).


15 Jul 06:

- Revamped the entire linear algebra section.  Routines now
  available as separate library.  Added TakagiFactor.


10 Apr 06:

- The SquaredME subroutine no longer sets the "reset" flag to
  .FALSE. after initialization.  Rather, this is done in xsection.F.
  This makes it easier to combine several squaredme modules.


7 Apr 06:

- Moved the Li2 function to LoopTools, substituted the Kronecker
  Delta function by a preprocessor macro.

- Updated configure to look for FeynHiggs and LHAPDF and also
  check more places.


27 Feb 06:

- Added the MultiplyDiagrams function of which the TagDiagrams
  function introduced earlier is just a special case.


20 Jan 06:

- Added the $ReadFormDebug variable.  When set to a non-zero value,
  ReadForm echos each line of the FORM output.


15 Dec 05:

- Added the Pool function which collects terms with common
  prefactors.


1 Dec 05:

- Added the OnSize function, which constructs a Switch
  statement for simplifying an expression according to its
  LeafCount.


23 Nov 05:

- Added the Abbreviate function, which introduces abbreviations
  for certain subexpressions (currently only sums).  Subexpr[]
  returns the definitions of the subexpressions and ClearSubexpr[]
  removes them.


25 Oct 05:

- The last change is now automatic with ./configure --quad.


20 Oct 05:

- Replaced all dble -> DBLE, dimag -> DIMAG, dconjg -> DCONJG.
  This way, the entire code can easily be treated to quadruple
  precision using the change in the makefile:
  FC = ifort -r16 -DDBLE=QEXT -DDIMAG=QIMAG -DDCONJG=QCONJG


21 Sep 05:

- Fixed bug re summation of Deltas in CalcFeynAmp.frm.

- Fixed problem with #call Shortest if args included parentheses,
  e.g. #call Shortest(..., MSf2(...)).


2 Aug 05:

- Corrected bug in PolarizationSum for massive vector bosons.
  Terms not containing the polarization vector were multiplied
  with 2; the correct number of polarizations is of course 3.


29 Jul 05:

- Added a generational dependence to the preprocessor macros
  for the sfermion mass-parameters, e.g. MLE -> MLE(g).


24 Jun 05:

- Improved internal InvSimplify FORM procedure which had some
  problems with large expressions.


8 Jun 05:

- Renamed internal symbols i$, d$, e$, p$, dummy$ to
  iJ, dJ, ... because Mathematica uses the $ also in Modules
  and then gives warnings of the kind "Form`i$ may be shadowed
  by Global`i$".


13 Apr 05:

- Consistently renamed all remaining occurrences of Mandel* to
  Inv*, e.g. MandelSimplify -> InvSimplify.


7 Mar 05:

- Fixed problem with insertions in Particle-level amplitudes.

- Updated some Cuba bug fixes.


9 Feb 05:

- Switched to unitary representation for the CKM matrix in sm_ini.F.


8 Feb 05:

- Added UASf + MASf stuff for FVMSSM.

- Added TagDiagrams function.

- Updated to Cuba 1.2.


7 Feb 05:

- Added cuba.F to the drivers.  Takes care of numerical integration
  using the parameters defined in run.F with a simple interface.

- Removed ModelDigest from *_ini.F files.  ModelVarIni now dumps
  the parameters once after each ModelConstIni invocation.
  Signalling done via the *_digest logical variables in model.h.


2 Feb 05:

- Added CompAZ to the util library.


25 Jan 05:

- Extended mssm_ini.F for linking with FeynHiggs 2.2.

- Updated to Cuba 1.1.

- Improved ReadData for reading huge data files.

- Log-file name now a 7-digit number.


12 Aug 04:

- Significant changes in the FORM part: Rational exponents now
  treated properly in FORM.  Scaling (AbbrScale) works properly.

- Improvements in PolarizationSum.

- Added InsertBefore option with which the loop integrals can
  be completely symmetrized.


18 Jun 04:

- Rewrote RemoveRedundancy, it now sorts tmp variables in such that
  they come after everything they depend on has been calculated
  (essentially like OnePassOrder applied to the result).


6 May 04:

- Added sqrtSstep to the output file name.

- Added submit script to automatically distribute run.F jobs over a
  cluster.


15 Apr 04:

- Added reorder utility which works much like restack in older versions.


14 Apr 04:

- openlog skips already existing logfiles.  This is useful if the
  executable dies or is killed.  One may need to remove the last,
  partial, log file, but can otherwise restart the program with
  the same parameters to resume the calculation.


9 Apr 04:

- Fixed some technical Fortran issues wrt g77 in squared_me.F.

- Added sqrtSstep in main.F.


8 Apr 04: lots of new stuff!

- Integrated the Cuba library [hep-ph/0404043] into util.a.

- Rearranged main.F so that the actual main program scans only the
  command line and then invokes the subroutine ParameterScan, which
  performs the actual calculation.

- Introduced serial numbers on the command line.  With this concept
  one can with very little effort parallelize a parameter scan on a
  large cluster.

- Every pass of the parameter-scan loop (LOOP1..15) yields an individual
  output file in all output is stored, messages and "true" data alike,
  where the latter are merely tagged by a @ in the first position.
  These output files are written to a directory named much like the
  output files in older versions, e.g. run.UUUU.00500/.  The script
  "data" extracts the actual data from the output files in this directory.

- Likewise, the main program yields as an exit value for the shell the
  number of passes of the parameter loop which were actually computed.

- The PARALLEL stuff has been removed.  Forking is much too valuable
  to waste on the simple parallelization that can be performed more
  efficiently using the serial numbers.  Forking is reserved for the
  parallelization of the integration algorithms.

- Returned to the SymbolPrefix option (rather than the $SymbolPrefix
  global variable).  Now implemented as PREFIX preprocessor def with
  the SymbolPrefix value as default.

- Helicities are now encoded as bit string.  Got rid of that stupid
  TrueHel this way.  Thus
    7 = 111b = use pols +1,0,-1 (unpolarized),
    2 = 010b = use only 0 (longitudinal),
    5 = 101b = use +1,-1 (transverse).
  Bits for particle 1 are at the lower end of the word.

- IDENTICALFACTOR is now multiplied also with the differential
  cross-section.  The factor comes from Bose statistics and not
  from phase-space overcounting and will hence be observed in
  differential cross-sections too.

- Archive now unpacks into subdirectory, no longer "flat."


27 Oct 03:

- Put the $SymbolPrefix in a lot of places where it was missing.


20 Oct 03:

- Enlarged TERMBUF limit in ReadForm.tm.


7 Oct 03:

- Rearranged some internal context stuff.

- Fixed some minor problems with PolarizationSum.


26 Jun 03:

- Moved "call setdelta" in main.F etc. into loop over parameters as
  DELTA, MUDIM, and LAMBDA may depend on model parameters.


3 Jun 03:

- Changed default to FermionChains -> Weyl in CalcFeynAmp.
  Rearranged all examples accordingly.


2 Jun 03:

- Got the Weyl-van der Waerden spinors going!
  Significant gains in both numerical stability and speed.

- When using standard Dirac fermions and the trace technique (i.e.
  HelicityME), the DIRACFERMIONS flag in process.h should be defined to
  compute unpolarized cross-sections efficiently, i.e. without actually
  summing up the different helicities.


27 May 03:

- Added PolarizationSum for performing the sum over the polarizations
  of the squared amplitude analytically.  The gauge-dependent terms
  (for massless external gauge bosons) are expressed in terms of a
  vector eta, but can also be turned off with GaugeTerms -> False
  (careful!).

- Moved all programs internal to FormCalc to the FormCalc subdirectory,
  much like in FeynArts.


30 Apr 03:

- Rearranged entirely the representation of vectors in the driver
  programs, now all functions in util/kin use the Weyl-van der Waerden
  representation, improves stability and is a first step towards the
  full Weyl spinor formalism.

- Removed VADecompose option of CalcFeynAmp.  It now has the new option
  FermionChains -> Chiral|VA|Weyl, which subsumes the functionality of
  VADecompose.  With FermionChains -> Weyl, Weyl spinor chains are
  returned by CalcFeynAmp.  Currently these are not useful for Fortran
  code generation.

- The renormalization constants are now split into several files if
  they become too large.


31 Mar 03:

- Fixed problem with phase in SingularValues, formerly
  U A V^+ = diag(d), now U^* A V^+ = diag(d).

- Rewrote Eigensystem and Eigen2x2 from scratch using Jacobi algorithm,
  much better precision now.


7 Mar 03:

- Added possibility to calculate amplitudes at Generic level.
  The generic objects not suitable for FORM, like the generic couplings
  (G[_][_][__][__]) are substituted by symbols, e.g. Coupling1.
  These symbols remain after CalcFeynAmp.  GenericList[] returns
  a list of substitutions, thus "result /. GenericList[]" recovers
  the original objects.


6 Mar 03: major rearrangement of the structure of the generated code
(version number therefore now 4.0 beta):

- New syntax for WriteSquaredME and WriteRenConst:
	dir = SetupCodeDir["myfortran", Drivers -> "mydrivers"];
	WriteSquaredME[..., dir];
	WriteRenConst[..., dir]
  i.e. setting up the code directory is now a separate step.	

- Generated modules are now put in own subdirectories (squared_me/ and
  renconst/ by default, may be changed with the Folder option of
  WriteSquaredME and WriteRenConst).  Only a minimalistic makefile is
  generated in each subdirectory.  FormCalc is now completely unaware
  of the drivers (they used to be mentioned as dependencies in the
  makefile).  Code generated by WriteSquaredME and WriteRenConst is
  completely independent of each other (option RenConstFile gone).
  The connection which renconst to use with which squared_me is now
  made by the central makefile.in in the drivers.

- The makefile.in in the drivers directory determines which
  subdirectories are taken into account.  This makes it easy to have,
  e.g., several subdirectories with squared matrix elements, such as
  for up- and down-type-quark processes when considering proton
  scattering.  Moreover, the makefile.in can be modified and copied
  to the custom driver directory since it is no longer overwritten by
  WriteSquaredME.

- WriteSquaredME's SymbolPrefix option replaced by the global
  variable $SymbolPrefix which is used by WriteSquaredME, WriteRenConst,
  SubroutineDecl, and CommonDecl.

- Added sfx shell script that packs all files necessary to compile
  the code generated by FormCalc into a self-extracting archive for
  easy distribution to other people.

- Added turnoff shell script with which the calculation of certain
  modules in squared_me.F can be turned off (i.e. the calls to the
  modules are commented out in squared_me.F).

- Moved pnuglot script to the drivers directory, making it unnecessary
  to set path to ~/FormCalc/tools.

- Much less dependence on preprocessor variables.  Most utility
  functions now moved to a library util.a which needs to be compiled
  only once.

- Main program now in main.F, the 1to2, 2to2, and 2to3 drivers now
  contain only the output routines for differential and integrated
  cross-sections, DoDifferential and DoIntegrated.

- Added one-dimensional integration algorithm "Patterson" which,
  though non-adaptive, gives an error estimate together with the
  integral.  The basic integration rule has 21 points; if the error
  is deemed too large, a 43-point rule is used (re-using all 21
  previously computed points); if that is still insufficient, a
  81-point rule is taken (again re-using all previous values).

- The executables run etc. are now invoked as "./run uuuu ..."
  instead of "./run u u u u ..." as before.  The new version makes
  it easier to scan the command-line arguments.

- The generated subroutine SquaredME (and indeed all other driver
  programs, too) no longer pass the results in two scalar variables
  tree and loop, but now in an array with NCOMP components (NCOMP is
  defined in run.F), where result(1) = tree and result(2) = loop.
  This makes it easy to add further components, e.g. if the square
  of the one-loop matrix element shall be returned, too.

- The third column in the result files, which used to be the
  tree-level result plus the one-loop contribution, is now only the
  one-loop contribution.  This way there is no loss of precision
  if the tree-level result is much larger than the one-loop
  contribution.


17 Feb 03:

- Made DCUHRE the default integrator for 2 -> 3.

- dcuhre.f now included by 2to3.F, obviating the need to explicitly
  compile libdcuhre.a.


12 Feb 03:

- Added OffShell function to take individual particles off shell.

- Removed last remaining CERNlib dependence by adding own routine to
  compute alpha_s.

- Added "--with-cernlib" switch to configure to restore old behaviour
  (e.g. alpha_s should be computed by PDFlib when using the parton
  distribution functions from PDFlib).


28 Jan 03:

- Rearranged much of the Gauss-Kronrod integration in multigauss to
  fix a bug. 

- Fixed up minor problem with allocation of working space in DCUHRE, too.


22 Jan 03: FormCalc 3.2 released, lots of new stuff:

- Completely general and systematic introduction of invariants
  for any number of external legs.

- Added an absolute precision goal to Vegas and Multigauss.

- Improvements in the ScanGraphics package.

- Improvements in ReadData (much faster now).

- 1 -> 2 drivers added.

- Added possibility for parallel execution of FormCalc-generated
  code (compile with -DPARALLEL, see hep-ph/0210220).

- Improvements in the code generation:
  1. Now uses preprocessor macros k(n), e(n), ec(n), and s(n) instead of
     integers to access vectors.
  2. New sumup function improves readability of squared_me.F considerably.
  3. Helicities passed to SquaredME as arrays, not single integers.
  4. sqrtS no longer passed to SquaredME.
  5. Because of the changes in the calling conventions (#3 & #4), the
     subroutine's name has been changed from squared_me to SquaredME
     to alert the user to the new conventions.

- Internal reordering mostly in the FORM part:
  1. Better partial factorization of expressions
  2. Simplification of invariants now done in FORM

- Needs Mathematica 4 or higher now.

- Compiled executables now reside in a single directory, which is named
  by Mathematica's $SystemID, rather than by the shell's $HOSTTYPE,
  e.g. Linux/{ReadForm,ToForm,ToFortran,ReadData}.
  ($HOSTTYPE had led to some complications because tcsh and bash had
  different values for it on some OSs.)

- mssm_ini now uses At, Ab, and Atau (rather than Au and Ad).
  Names have deliberately been chosen different than before because
  Atau must be given in addition to At and Ab, hence this is not
  just a renaming, but there must be at least a statement "Atau = Ab"
  or similar.

- mssm_ini has its own set of SVD and diagonalization routines and is
  hence no longer dependent on LAPACK (libmathlib.a of CERNlib).
  The conventions are now that all mass arrays are sorted into ascending 
  order (MNeu and MCha were in decreasing order so far).  This matches
  the conventions commonly adapted in the literature.


13 Sep 01:

- Added adaptive Gauss-Kronrod integration for 2 -> 3 (multigauss.F),
  algorithm can be tuned using the NODES_var preprocessor defines in run.F
  and may be significantly faster for some applications, though also worse
  for nasty integrands.

- Added special cases of B1[0, ...] to tools/btensor.m.


30 Aug 01:

- Largely rewritten the internal FormCalc functions for optimization and
  pre-output ordering: OptimizeAbbr, Dependencies, and OnePassOrder, yields
  some 30% in speed.


27 Aug 01:

- Switched 2 -> 3 code from (fixed-point) multiple Gaussian integration to
  VEGAS, added the adaptive ACM-algorithm DCUHRE for checking.