Beta4.9.4-1.txt

                  Geant4 9.4-beta-01 Release Notes
                  --------------------------------

                                                            25 June 2010

List of main features and fixes included in this Beta release since 9.3.p01:

  o Configuration:
    -------------
    + Configure script:
      o Protected echos on LD_LIBRARY_PATH so that output from use of
        Configure post-install is clean.
      o Added removal of quotes from make version string to handle non-English
        language systems.
      o Fixed selection of XERCESCROOT path for Windows.
      o Added new data set G4NEUTRONXS (Evaluated neutron cross section
        data, version 1.0), controlled by variable G4NEUTRONXSDATA.
      o Updated data set G4EMLOW to version 6.13.
    + Removed 'leading_particle' hadronic module from include paths list.
    + Increased stack size on WIN32-VC setup (i.e. 8MB as on Linux), adding
      the option to LDFLAGS.

  o Digits+Hits:
    -----------
    + Added missing virtual destructor to G4VScoringMesh and removed
      unnecessary virtual qualifiers to G4ScoringBox.

  o Environments:
    ------------
    + G4Py:
      o Revised usage of MultipleScattering classes to make use of new particle
        based processes.
      o Fixed bug in wrapping of G4GDMLParser.
      o Updated run-manager and physics-lists.
    + Momo:
      o Revised obsolete trajectory drawing.

  o Event:
    -----
    + Improved logic in G4PrimaryTransformer. Addressing problem report #1101.
    + Fixed compilation warning in G4SmartTrackStack issued with gcc 4.5.0.
    + Added few accessors methods to G4SPSEneDistribution class.
    + Removed obsolete CLHEP_HepMC option from G4EventManager class.

  o Geometry:
    --------
    + solids/Boolean:
      o Introduced recursive algorithm in CreatePolyhedron() for Boolean
        operations: it uses HepPolyhedronProcessor from 'graphics_reps' module,
        using new technique in attempt to avoid numerical problems for the
        calculation of the polyhedron in BooleanProcessor. It allows to try
        all permutations, also for Booleans of Booleans.
        Helps in reducing the number of cases of "Error in Boolean processor"
        for visualization; still some stubborn cases are left.
    + solids/CSG:
      o G4Box: first implementation of speed improvements and corrections from
        joint code review of G4Box class: introduced half-tolerance constants
        and better logic in some conditional statements. Some code cleanup.
      o G4Orb: modified DistanceToIn(p,v) to be more consistent with Inside(p).
    + solids/specific:
      o Introducing new G4GenericTrap shape, a new solid representing an
        arbitrary trapezoid with up to 8 vertices standing on two parallel
        planes perpendicular to the Z axis.
      o Fix in G4TriangularFacet and G4TessellatedSolid to correct treatment of
        optical photon transport related to internal reflection at surface.
        Addresses problem report #1103.
      o G4ExtrudedSolid: eliminated requirement for clockwise ordering of
        polygon vertices. Added a check for vertices ordering; if vertices
        are defined anti-clockwise their ordering is reverted.
        Fix in polygon facet triangularization for consequent concave vertices.
      o Fixed deletion of internal arrays in CreatePolyhedron() for G4Polycone
        and G4Polyhedra.
    + volumes:
      o Make use of specialized allocator for handling internal vector in
        G4NavigatorHistory, globally controlling the memory pool, to optimise
        memory management and reduce fragmentation.
        Measured ~2% average run-time speed-up.
      o Added Reset() method to G4ReflectionFactory for clearing maps of
        constituent and reflected volumes.

  o Global:
    ------
    + Introduced G4PhysicsVectorCache, placeholder for cache R/W data in use
      by G4PhysicsVector and derivates. Modified G4PhysicsVector and
      G4LPhysicsFreeVector accordingly.
    + Moved interpolation methods to inline in G4PhysicsVector.
    + More safe numerical computation of A13() and logA() methods in G4Pow.
    + Changed date for release 9.4-beta.

  o Graphical Representations:
    -------------------------
    + Added support for new command "/vis/scene/add/digits", analogous to the
      already-existing "/vis/scene/add/hits": added methods Draw() and
      FilterDigi() for digits in G4VVisManager; added method AddCompound ()
      in G4VGraphicsScene.
    + Added method DispatchToModel(const G4VTrajectory&), i.e., without i_mode,
      to G4VVisManager in order to distinguish usage and in preparation for
      'i_mode' migration.

  o Intercoms:
    ---------
    + G4UIManager: added field to get GUI Window session.
    + Added used characters to 'tokenNum' in G4UItokenNum suppressing compiler
      warnings of gcc-4.5.0.

  o Interfaces:
    ----------
    + G4UIExective: changed order for session type; GUI session is prioritized.
    + G4VBasicShell: implemented improvements for command completion.
    + G4Qt: added a link to G4UIQt
    + G4UIQt: added tab widgets for visualization. Added new command in help
      tree for immediate update of the tree. Fixed warning in standard-output
      during viewer creation and fixed flush problem. Fixed recursive repaint
      problem. Fixed call to contructor QKeySequence() in AddButton() to
      allow for porting on gcc-4.5.0.

  o Materials:
    ---------
    + G4MaterialPropertyVector: removed allocation of temporary helper object
      and modified GetAdjacentBins() to allow for a value to match a bin
      exactly. Measured 21% CPU speed-up for optical processes.
    + G4OpticalSurface: changed GetAngularDistributionValue(), no longer
      dereferencing a pointer but returning the value at the array index.
    + G4NistMaterialBuilder: added 6 new materials.
    + G4NistMaterialBuilder, G4NistManager: added method
      GetMeanIonisationEnergy(). Added material G4_LUCITE.
    + G4ionisParamMat, G4DensityEffectData: added possibility to access density
      effect data in the case when pure material is not defined via NIST; atom
      number and state of such materials are checked.
    + G4DensityEffectData: fixed method GetIndex(Z, state) for Hydrogen
      and solid states, renamed it to GetElementIndex().
    + G4IonisParamMat: added member and method GetInvA23(); use
      GetElementIndex().
    + G4IonisParamElm: use mean excitation energy from NIST DB; addressing a
      problem reported for mean energy deposition in compound materials.
    + G4AtomicShells: improved performance by adding indexing array.
    + G4Element: moved include of G4AtomicShells header to the source.

  o Particles:
    ---------
    + Allow creation of shortlived particles other than in 'PreInit' state.
    + ElectronOccupancy in G4DynamicParticle is now created on demand, i.e.
      it is not created at construction of G4DynamicParticle but when
      AddElectron() or RemoveElectron() are invoked.
    + Fixed problems that error messages appear in cout and not in cerr.
    + Added anti-nuclei as short-lived particles.
    + G4MuonRadiativeDecayChannelWithSpin: corrected sign error.
    + Code cleanup in G4DynamicParticle constructors.

  o Persistency:
    -----------
    + Added G4GenericTrap among the set of supported solids in both GDML reader
      and writer, treating as 'arb8' GDML type.
    + Corrected dump of division axis in GDML writer.
    + ASCII module: added missing virtual destructor to G4tgbDetectorBuilder.

  o Physics Lists:
    -------------
    + Moved QGSP_FTFP_BERT list to supported (was experimental).
    + Removed previously blocked lists (FTFP_EMV, QGSC_EFLOW, QGSP_EMX).
    + Remove obsolete lists, except QGSP (now blocked) and QGSC_CHIPS;
      i.e. lists removed: FTFC, FTFP, LHEP_BERT_HP, LHEP_BERT, LHEP_PRECO_HP,
           QGSC_EMV, QGSC, QGSC_QGSC, QGSP_BERT_DIF, QGSP_BERT_NQE,
           QGSP_DIF, QGSP_EMV, QGSP_EMV_NQE, QGSP_NQE, QGSP_QE
    + New physics list QGSP_BERT_CHIPS: use CHIPS inelastic for all "misc"
      particles, including anti-proton, anti-neutron and hyperons.
      Also use interface to CHIPS cross-sections for Kaons.
      Added classes: QGSP_BERT_CHIPS, HadronPhysicsQGSP_BERT_CHIPS, and
      G4MiscCHIPSBuilder; modifed G4PhysListFactory accordingly.
    + Introduced new constructor for all builders with verbosity argument.
      Existing constructors are kept, but updated default arguments to avoid
      ambigous signatures. Changed physics-lists to use the new constructor.
      Introduced new constructor for G4NeutronTrackingCut.
    + Replaced G4HadronQElasticPhysics with G4QElasticPhysics and G4IonPhysics
      with G4QIonPhysics in the CHIPS physics-list 
      Bug fix for the virtual destructor in G4QElasticPhysics & G4QIonPhysics.
      Particles are restricted to SU(3) hadrons (no pi0 & eta) in
      G4QElasticPhysics.
      Fix in G4QPhotoNuclearPhysics to reset particles iterator.
    + G4EmStandardPhysics_option2: added G4WentzelVIMscModel and 'RangeFactor'
      set to 0.04 for e+-.
    + G4EmLivermorePhysics, G4EmPenelopePhysics, G4EmDNAPhysics,
      G4EmLivermorePolarizedPhysics: added G4GoudsmitSaundersonMscModel for e+-;
      added 1 MeV upper limit for the Livermore ionisation model (was 1 GeV).
    + Revised QGSP_FTFP_BERT and FTFP_BERT lists to use CHIPS G4QInelastic for
      all "misc" particles, including anti-proton, anti-neutron  and hyperons,
      and use interface to CHIPS cross-sections G4QHadronInelasticDataSet for
      kaons. Required changes to respective HadronPhysics* classes.     
    + LBE: substituted obsolete G4MultipleScattering by particle wise processes.
    + QBBC: disabled multi-fragmentation model of deexcitation.
    + G4FTFPNeutronBuilder, G4FTFPPiKBuilder, G4FTFPProtonBuilder,
      G4QGSPNeutronBuilder, G4QGSPPiKBuilder, G4QGSPProtonBuilder, G4QGSBuilder:
      added deletion of G4ExcitationHandler and G4QGSMFragmentation; added
      deletion of neutron builder for QGSP_BERT, FTFP_BERT, QGSP_FTFP_BERT.

  o Electromagnetic Processes:
    -------------------------
    + Adjoint:
      o G4AdjointhMultipleScattering: removed unused header.
    + High Energy:
      o New model G4ICRU73NoDeltaModel, derived from G4ICRU73QOModel.
    + Low Energy:
      o Added four new models (gamma ray models Compton, GammaConversion,
        PhotoElectric, Rayleigh), upgrades of G4Penelope from version 2001 to 
        version 2008.
      o Added MinEnergyCut() method to G4PenelopeIonisationModel. Small (minor)
        changes to other Penelope models (mainly clean-up).
        Removed useless call in G4PenelopeIonisationModel which triggered fake
        warning messages.
      o Updated public interface of G4LivermorePhotoElectricModel.
        Removed possibility to use custom cuts for fluorescence/Auger.
        Main cuts from G4ProductionCutsTable are always used, for consistency.
      o New G4PenelopeOscillator* classes (Beta version) to manage
        multi-element atomic oscillators used in Penelope Compton and
        Ionisation models. Compliant with Penelope-2008.
      o Added dummy method ComputeCrossSectionPerAtom() in
        G4PenelopeRayleighModel, never inkoved in tracking and issuing a
        warning if users access it via G4EmCalculator.
      o Clean-up in Penelope models; explicitely set Auger flag to false in 
        constructor.
      o Fixed initialisation order in some models' constructor to fix Valgrind 
        run-time errors reports.
      o Fixed bugs in G4AtomicDeexcitation and G4AugerTransition.
      o G4DNA processes:
        - Added new prototype excitation model for electrons in
          G4DNABornExcitationModel. Will be fully released in December.
        - Corrected G4DNABornIonisationModel for maximum kinetic energy
          transfer.
        - Modified method declaration in G4DNABornExcitation. Modified electron
          correction in G4DNAMillerGreenExcitation.
        - Extended search range for maximum of DCS in G4DNARuddIonisationModel.
        - Changed default parameters in G4DNAEmfietzoglouExcitationModel.
          Modified initialisation of e- Emfietzoglou excitation model to fix
          memory leaks.
        - Changed computation of scattering angle in G4DNAChampionElasticModel.
        - Replaced hard coded masses in G4DNADingfelderCharge* classes.
      o New PIXE developments:
        - G4ecpssrKCrossSection and G4ecpssrLiCrossSection classes: these
          classes substitute the G4ecpssrCrossSection class, fixing existing
          bugs and improving the goodness of the results especially at higher
          energies.
          The Geant4 PIXE developers are really thankful to Maria Grazia Pia
          for searching for bugs in the previous Geant4 release, pointing
          existing problems and propose solution in a very collaborative way
          in: M.G.Pia et al., CHEP-2009 conference, March 2009, Prague (Czech
          Republic).
        - G4empCrossSection and G4teoCrossSection classes: these two classes,
          deriving from G4VhShellCrossSection, provide the interface to
          empirical ionization cross sections (Orlic for L shells and Paul
          for K).
        - G4hShellCrossSection, G4hShellCrossSectionDoubleExp,
          G4hShellCrossSectionDoubleExpData, G4hShellCrossSectionExp,
          G4hShellCrossSectionExpData classes are now discontinued.
        - G4hLowEnergyIonisation has been updated to follow the new design
          and let the user chose which model to use for the calculation of shell
          ionization cross section through the method SelectShellIonisationCS().
          Dependencies on old methods for cross sectiuon calculation have been
          removed.
        - Bugs fixes in G4OrlicLiCrossSection and G4PaulKCrossSection.
      o Fix in G4AugerTransition to cure crash on Windows.
      o Added missing virtual destructors where necessary.
      o Corrected return value case in method GetPhotoElectronDirection()
        in G4PhotoElectricAngularGeneratorSauterGavrila.
        Addressing problem report #1120.
      o Requires new G4LEDATA data set, version 6.13.
    + Muons:
      o G4MuPairProductionModel: added sampling recoil of a primary particle
      o G4MuIonisation: use G4ICRU73QOModel for mu- for E< 0.2 MeV.
    + Polarisation:
      o Fix in G4ePolarizedIonisation to use default vector size of
        G4VEnergyLossProcess and avoid size mismatch in G4LossTableBuilder.
    + Standard:
      o G4GoudsmitSaundersonMscModel: fixed problems of small angle theta
        distributions. Improved numerical stability of computations for small
        angles by direct cross-section calculation, now inverse of the inverse
        angular sampling without large angle rejection method longitudinal
        displacement is computed exactly from <z>.
      o G4GoudsmitSaundersonTable: improved numerical stability by minimum of
        used arrays in computation within the dichotomy finding method.
        Optimized code to improve speed by addition extra class members.
      o G4eBremsstrahlungModel: reduced 'minThreshold' to 0.1 keV.
      o G4MollerBhabhaModel: reduced minimum cut value to mean ionisation
        potential.
      o G4WentzelVIModel: added protections; added DefineMaterial() call to
        cross-section class.
      o G4ASTARStopping, G4PSTARTStopping: replaced C-arrays with
        G4PhysicsVector with Spline interpolation. Reviewed names of materials
        (G4_PHOTO_EMULSION, G4_He, G4_NYLON-6-6).
      o Fixed G4_O data in G4ASTARStopping. Fixed G4_Pt and G4_PROPANE data in
        G4PSTARStopping.
      o Cleanup in GetElectronicDEDX() method in G4WaterStopping.
      o G4eMultipleScattering, G4hMultipleScattering: improved printout at
        initialisation. Removed old unused header of G4UrbanMscModel in
        G4hMultipleScattering.
      o G4eBremsstrahlungRelModel, G4MollerBhabhaModel: set minimum cut value
        to 0.1 keV.
      o G4eBremsstrahlungModel: fix in SelectRandomAtom() method.
        Addressing problem report #1115.
      o G4UrbanMscModel93: added randomization of step limit for the first
        step in any volume for option "UseDistanceTo Boundary".
      o G4BraggIonGasModel, G4BetheBlochIonGasModel: new ionisation models for
        low density media.
      o G4BraggModel: above 2 MeV use ICRU'49 parameterisation for all atoms.
      o New model G4ICRU73QOModel for anti-particles
      o G4hIonisation: use new G4ICRU73QOModel for negatively charged particles
        instead of G4BraggModel, which used to be applied for positevely
        charged particles only. This modification provides simulation of
        Barkas effect at low energies (before the effect was simulated at
        higher energies in G4BetheBlochModel).
        The high energy limit for the new model is the same as for G4BraggModel
        (2 MeV for protons/anti-protons).
      o G4WentzelOKandVIxSection: new class to compute cross sections and
        sample scattering angle.
      o G4WentzelVIModel, G4eCoulombScatteringModel, G4CoulombScatteringModel:
        use new class G4WentzelOKandVIxSection.
      o G4WentzelVIModel: optimized method of step limitation.
      o G4WentzelVIModel, G4eCoulombScatteringModel: added momentum dependent
        limit of scattering angle between single and multiple scattering which
        allows to use G4WentzelVIModel with optical model of hadron scattering.
      o G4PAIModel, G4PAIPhotonModel: added printout for the model
        initialisation.
      o Removed obsolete models: G4UrbanMscModel, G4UrbanMscModel2,
        G4UrbanMscModel71, G4MultipleScattering, G4ComptonScattering52,
        G4PolarizedComptonScattering.
    + Utils:
      o G4VMscModel, G4VMultipleScattering: set skin=1.0 as default.
      o G4LossTableManager: added new method GetNumberOfBinsPerDecade(), which
        is used in G4VEmModel for initialisation of G4EmElementSelector.
        Addressing problem report #1115.
        Added methods PreparePhsyicsTables(), BuildPhysicsTables(), fixing
        initialisation of models via G4EmConfigurator.
      o G4EmProcessOptions: removed extra initialisation code, already existing
        in G4LossTableManager.
      o G4VEnergyLossProcess, G4VEmProcess, G4VMultipleScattering: added
        calls to new G4LossTableManager methods PreparePhsyicsTables() and
        BuildPhysicsTables().
      o Added method GetCurrentElement() in G4VEmProcess, G4VEnergyLossProcess
        and G4VEmModel classes.
      o G4VEmProcess: cleanup printout at initialisation for scattering process.
      o G4VEmModel: added method ChargeSquareRatio() to access current charge
        of an ion; use this method in G4VEnergyLossProcess.
      o G4VEnergyLossProcess: call CorrectionsAlongStep() only for ions (minor
        CPU saving). Use the same method to build cross section table
        as DEDX table (use copy constructors to reduce number of calls to exp()).
      o G4EmCalculator: fix in GetCrossSection() method.
      o G4EmElementSelector: use Spline flag to construct vectors probabilities.
      o G4EmModelManager: do not use min energy cut defined by models, so
        production threshold is defined only by range to energy conversion and
        is limited only by low limit of energy threshold defined in 'cut'
        category.
      o G4EmConfigurator: fix for case of more than one model added per a
        process.
      o G4EmCorrections: moved G4AtomicShell header into source.
    + Xrays:
      o Included scintillation rise time to G4Scintillation.
      o Use std namespace for mathematical functions explicitly.

  o Hadronic Processes:
    ------------------
    + Removed obsolete code for 'leading_particle' model.
    + cross_sections
      o G4NeutronElasticXS, G4NeutronInelasticXS, G4NeutronCaptureXS: added
        path to new data set through environment variable G4NEUTRONXSDATA.
      o Implemented integer Z and A transition to G4CrossSectionDataStore.
      o G4UPiNuclearCrossSection: fixed memory leak at destruction.
      o Updated G4GlauberGribovCrossSection for anti-protons.
    + management
      o G4HadronicProcess: added methods CheckEnergyMomentumConservation(),
        SetEpReportLevel(), SetEnergyMomentumCheckLevels() and
        GetEnergyMomentumCheckLevels(), which implement energy/momentum 
        checking. Some code cleanup.
      o New G4HadronicEPTestMessenger class to provide energy/momentum test 
        commands available in UI.
    + models/cascade
      o Replaced all uses of bindingEnergy() (local Bertini method) with the
        Geant4 standard G4NucleiProperties::GetBindingEnergy(A,Z) in classes
        G4IntraNuclearCascader, G4InuclCollider, G4NucleiModel, G4Fissioner,
        G4EquilibriumModel and G4NonEquilibriumModel.
        G4Fissioner still uses Bertini method bindingEnergyAsymptotic().
      o G4Diproton, G4Dineutron, G4UnboundPN: new G4ParticleDefinition
        subclasses, which implement the Bertini-specific dibaryon states.
      o G4ElementaryParticleCollider: replaced original (incorrect) pp, pn, nn
        2-body to 2-body scattering angular distributions with a new
        parameterization of elastic scattering data using the sum of two
        exponentials.
      o Improved angular distributions for 2-body final states of pi-nucleon,
        nucleon-nucleon, kaon-nucleon anf hyperon-nucleon scattering.
      o Replaced G4CascadeMomentum internal class with G4LorentzVector.
      o Completed first stage of migration from using internal Bertini particle
        types to Geant4 standard particle types.
      o Energy/momentum checks added which remove large energy non-conservation
        events.
      o Improved memory management: removed allocation of strings in various
        constuctors; replaced several cases of return-by-value with const or
        non-const reference, or by use of buffers.
      o CPU speed improvements from the removal of unnecessary class
        instantiations.. Code re-organization to permit more speed improvements
        in the future.
    + models/chiral_inv_phase_space
      o Updated proton elastic and CHIPS parameters.
      o Implemented anti-baryon elastic interactions.
      o Added pion elastic and fixed a bug in CHIPS fragmentation.
      o Implemented isotope wise CHIPS elastic for neutrons.
      o Fix in G4QElasticCrossSection in parameter order for nH. 
      o Bug fix in G4QNeutronNuclearCrossSection for Ba isotopes.
      o Bug fix for K0 in G4QKaonMinusElasticCrossSection::GetExchangeT().
      o Bug fix in G4QNucleus::ChoosePosition() for looping.
      o Fix to use of Z+N instead of N for A.
      o Corrected Titanium and Zirconium isotopes in G4QIsotope.
      o Added decay of Chipolino in the G4QEnvironment, when A_env=0.
      o Bug fix following G4NucleiProperties warning for A=0, Z=0.
      o Bug fix in G4QPDGCode and G4QEnvironment for Omega and Sigma;
        when mass is small for Lambda andSigma+, use Xi0,p.
      o Correction in G4QElastic for K0 elastic cross-sections.
      o Corrected treatment of vacuum in fragmentation module.
      o nn is decayed in G4QLowEnergy; Omega- is included in G4QElastic;
        N neutrons (A=N, Z=0) are decayed in G4QLowEnergy.
      o Bug-fix in G4QLowEnergy: pions as secondaries.
      o Bug-fix in K0 elastic & Ion-Ion Elastic.
    + models/coherent_elastic
      o Use G4Q(Proton/Neutron)ElasticCrossSection instead of
        G4QElasticCrossSection; fixed initialization problem for
        G4QNeutronElasticCrossSection.
      o G4UHadronElasticProcess: fixed CHIPS warning for Cu, due to wrong
        isotope sected; addressing problem report from ALICE.
    + models/de_excitation
      o G4StatMFMacroTemperature: cleanup logic of solving equation for
        temperature; moved constructors and destructor to source.
      o G4ProtonEvaporationProbability, G4DeuteronEvaporationProbability,
        G4TritonEvaporationProbability, G4He3EvaporationProbability,
        G4AlphaEvaporationProbability: return back to published variant
        parameterization of inverse cross section (OPT3, Kalbach).
      o G4FermiFragmentsPool: extended list of stable fragments.
      o G4UnstableFragmentBreakUp: new class to decay exotic fragment
        (like 2n or 2p).
      o G4DiscreteGammaTransition: make transition depended on Z and A (before
        was only Z) and added energy tolerance
      o G4VGammaDeexcitation: rewritten kinematic of 2-body decay.
        Fix to take into account electron mass in the case of electron emission.
        Take into account bounding energy of electron in the case of electron
        emission; fixed kinematic.
      o G4DiscreteGammaTransition, G4DiscreteGammaDeexcitation, 
        G4ContinuumGammaDeexcitation: set energy tolerance 1 keV. More accurate
        Lorentz computations. Define destructors to be virtual.
      o G4DiscreteGammaTransition: removed unphysical corrections of gamma
        energy; fixed default particle as gamma; avoid subtracting binding
        energy in case of electron emission.
      o G4ExcitationHandler: apply FermiBreakUp to fragments with A>1 (before
        was A>4) in order to reduce number of fake gamma produced in
        deexcitation of light fragments; added parameter 'minExcitation' equal
        to 1 keV. Added check on stability of primary; do evaporation if
        FermiBreakUp cannot deexcite a fragment. Added SetParameters() method.
      o G4Evaporation: rewritten BreakUp() method; added Initialise() method at
        construction, to setup all options and not at run time.
        Added InitialiseEvaporation() method to setup decay channels; changed
        order of decay channels: first photon evaporation, then fission,
        finally all other channels as before. Improved condition on how to
        stop deexcitation loop.
        Added call to G4UnstableFragmentBreakUp if natural abundances of cold
        fragment is zero and Z < 20; optimized logic of stopping of evaporation
        loop. Set as a default variant evaporation combining standard plus GEM
        probabilities.
      o G4PhotonEvaporation: added correction of electron state after emission.
      o G4VEvaporationChannel, G4PhotonEvaporation: added two new virtual
        methods EmittedFragment() and BreakUpFragment().
      o G4FermiConfiguration: parameter of Coulomb energy Kappa is changed from
        1 to 6 according to recommendation of original author of the model
        A.Botvina.
      o G4FermiPhaseSpaceDecay: improved model of sampling of kinetic energy.
        Cleanup of the code to use G4Pow.
      o G4FermiFragmentsPool: fix for fragment 111.
      o G4StatMFFragment, G4CompetitiveFission, G4EvaporationProbability,
        G4GEMProbability: corrected inclusion of headers.
      o G4E1Probability: code cleanup and optimisation by usage of G4Pow,
        integer A and introduction of const members. Added numerical protection
        to avoid division by zero.
      o G4GEMProbability: fixed numerical problem (division by zero), code
        cleanup and optimisation by usage of G4Pow and integer Z,A.
      o Code cleanup. Moved constructors and destructors to source files for
        virtual interfaces. 
    + models/im_r_matrix
      o Removed empty source file for G4VScatterer.
        Synchronised developments to HEAD version.
    + models/incl
      o Nuclear deformation and shell effect data reader now constructs the
        corresponding data tables correctly.
      o Added some internal logging facilities to INCL (disabled by default).
      o Added missing virtual destructor to base classes where necessary.
    + models/management
      o G4HadronicInteraction: added two methods, GetEnergyMomentumCheckLevels()
        and SetEnergyMomentumCheckLevels() in order to implement checking
        of energy/momentum conservation.
      o G4VPreCompoundModel: moved constructor and destructor to source.
    + models/neutron_hp
      o Modified sampling in G4NeutronHPInelasticCompFS.
      o  Fix bug about incidence energy in G4NeutronHPEnAngCorrelation.
    + models/parton_string/diffraction
      o Added missing virtual destructor to G4DiffractiveHHScatterer.
    + models/parton_string/hadronization
      o Improved Lund string fragmentation.
      o Fixed inconsistency between LightFragmentationTest and Lund
        fragmentation. Fixed problems of energy non-conservation in FTF.
    + models/parton_string/qgsm
      o Remove unused class G4PartonStringAnnihilator.
    + models/pre_equilibrium
      o G4PreCompoundModel: moved constructor and destructor to source.
      o Removed unused dummy classes G4VPreCompoundIon and
        G4VPreCompoundNucleon. 
      o G4PreCompoundProton, G4PreCompoundDeuteron, G4PreCompoundTriton,
        G4PreCompoundHe3, G4PreCompoundAlpha: return back to published variant
        OPT3 (Kalbach) parameterization of inverse cross section.
    + models/radioactive_decay
      o G4RadioactiveDecay:
        - In LoadDecayTable() create a decay table for isomers not included
          in RDM database and assume they all under go IT decay.
        - After DoDecay() check if there is any secondary produced.
          Kill the track if not to prevent infinite loop.
        - Insert a special treatment for K-40 beta decay; applied SetICM(),
          SetARM() and SetHLThreshold() to a decay channel.
          Addresses problem report #1068.  
        - Make sure the propertime is positive; negative case occures when the
          isomer is not in RDM database and its f-l is set to -1 by default.  
        - Corrected typo in declaration.
      o G4NuclearDecayChannel: added initialisation to all three constructors. 
      o G4RadioactiveDecay, G4NuclearDecayChannel:
        - Limit the shell index to < 7, as required by the ARM. Changed to use
          BreakUp() rather than BreakItUp() so to limit to one transition per
          step when ICM is applied. Addresses problem report #1001.
        - Added private member data 'applyICM', 'applyARM' and
          'halflifethresold' with their modifiers.
      o G4RadioactiveDecayMessenger: added the UI commands for SetICM, SETARM
        and SetHLThreshold.
    + models/util
      o G4Fragment: do not add any shift to excitation energy but use direct
        computation (this is needed to avoid infinite loop in gamma
        deexcitation); use G4NucleiProperties to access or compute ground
        state mass. Allow e- as a fragment.
        Added inline method ComputeGroundStateMass(); added accessors for
        integer Z and A; added ExcitationEnergyWarning() method.
        Minor speedup by adding member and access method to GroundStateMass.
        Some code cleanup: single return in inline methods.
    + util
      o Added interface for integer A&Z to G4Nucleus, and modified G4LightMedia,
        G4HadronicWhiteBoard and G4ReactionDynamics to use the new interface.

  o Optical Processes:
    -----------------
    + Fix for backpainted surfaces to avoid applying twice reflectivity.
       Addressing problem report #1114.
    + Refashioned code in the case of type 'dielectric_dielect'.
    + Fixed sampling of 'theFacetNormal' for 'polishedbackpainted' material
      in G4OpBoundaryProcess::DielectricDielectric() method. Address problem
      reported on Hypernews Optical Processes Forum, thread #334.
    + Modified G4OpBoundaryProcess to call ProposeLocalEnergyDeposit() every
      time a photon is in 'fStopAndKill' state because of NoRINDEX. This will
      trigger scoring when a photon encounters a volume not defined as an
      optical medium.

  o Run
    ---
    + Corrected printing of number of events processed on run abortion.
    + Fixed definition of path for Windows in G4RunManager for random seeds
      directory: use '\' instead of '/' and suitable system command.
    + Use "const G4String&" as arguments and return value wherever applicable
      in G4RunManager and G4RunManagerKernel.

  o Tracking
    --------
    + G4VTrajectory, G4Trajectory, G4SmoothTrajectory: added deprecation
      message, being printed via G4Exception only once, and only if a user
      invokes with a non-zero value of 'i_mode' in DrawTrajectory().

  o Visualization:
    -------------
    + management:
      o Added new "/vis/open" command for OpenGL viewers, "/vis/open OGL".
        This new form allows the same macro to be used whether one is on
        Linux or Windows (calling OGLSX on Linux and OGLSWin32 on Windows).
        When the application has been built with Qt, it will use the Qt form
        of OpenGL; when built with Motif, it will use the Motif form of OpenGL.
        Also added options OGLI and OGLS to force Immediate or Stored modes.
        The user is still free to use the older, specific viewer names, such
        as OGLIX, but the simple form, OGL, will give the most portable macros.
      o Added new "/vis/open" command for OpenInventor, "/vis/open OI".
        This new form allows the same macro to be used whether one is on Linux
        or Windows (calling OIX on Linux and OIWin32 on Windows).
      o Added "/vis/scene/add/digis" and "/vis/filtering/digi" commands,
        analogous to the already-existing commands "/vis/scene/add/hits" and
        "/vis/filtering/hits". As with hits, it remains the user's
        responsibility to provide the digitisation classes.
        Added method AddCompound(const G4VDigi&) to drivers.
      o Introduced DispatchToModel without 'i_mode'. The archaic form
        G4VTrajectory::DrawTrajectory(i_mode) is replaced by the newer form,
        G4VTrajectory::DrawTrajectory(). 'i_mode' used to be a parameter to
        control whether the trajectory was drawn as a line, a line plus step
        points, or only step points:
        - i_mode > 0 meant draw lines and points, size of points in
                     pixels = abs(i_mode)/1000
        - i_mode = 0 meant draw only lines
        - i_mode < 0 meant draw only points, size of points in
                     pixels = abs(i_mode)/1000
        Since release 8.1, better control has been available from the command
        "/vis/modeling/trajectories". Since release 5.0, 'i_mode' less than
        1000 has in fact drawn no points and 'i_mode' of 1000 has drawn points
        that were so small (a single pixel wide) that they were overshadowed
        by the trajectory itself except at endpoints of the trajectory or
        where the trajectory made a sharp change of direction.
        Use of 'i_mode' is now trapped and a warning gets printed.
      o Changed order of flags to follow G4UIExecutive.
      o G4VisManager, G4VisExecutive: introduced optional verbosity string
        argument in constructor; made 'fVerbosity' and GetVerbosity() static.
      o Added virtual destructors to fix compilation warnings.
    + modeling:
      o Added G4DigiModel and G4DigiFilterFactories.
      o Added default colours to G4TrajectoryDrawByParticleID.
      o Replace polyhedra with solids for sections and cutaways.
      o Trap use of 'i_mode' and print warning.
    + HepRep:
      o Added command "/vis/heprep/renderCylAsPolygons".
    + gMocren:
      o Removed dependency on G4VisManager. Improved printout.
    + OpenGL:
      o Some improvements with auto rotation mode.
      o G4OpenGLQtViewer: better inclusion of new tabs in UI part.
        Many improvements to OpenGL Qt.
      o Speed improvements for polymarkers (e.g., trajectories).
      o Corrected strdup usage in G4OpenGLViewer.
      o Use generic section and cutaway algorithm.
      o Added protection against multiple repaint call, and protection against
        resizeGl with bad values.
    + OpenInventor:
      o Added protection against repeated initialisation.
    + RayTracer:
      o Added missing virtual destructor to G4VRTScanner, G4RTXScanner and
        G4RTSimpleScanner classes.

  o Data Sets:
    ---------
    + G4EMLOW-6.13:
      o Updated PDF.dat to msc_GS module.
      o New directory and data files needed by ionisation cross section
        models for PIXE; added data tables in the directories uf, kpcsPaul
        and kacsPaul included in the 'pixe' dir.
      o Added 'sigmadiff_cumulated_elastic_e_champion.dat' and
        temporary 'sigma_excitation_e_born.dat' to dna module.
      o Added data files in the penelope module, necessary for
        Penelope08 models for gamma rays.
    + G4NEUTRONXS-1.0:
      o New data set for Evaluated neutron cross section data on natural
        composition of elements.

  o Examples:
    --------
    + Various fixes, improvements, adoption of new features.
    + Updated visualization usage in most examples, including macro sections
      the user can uncomment to activate additional features.
    + Adopt G4UIExecutive to initialise UI sessions.
    + Migrated physics-lists to use particle-based multiple-scattering
      processes.
    + advanced/human_phantom
      o Added missing includes, causing compilation error.
      o Corrected commands in some macros.
    + extended/electromagnetic/TestEm0
      o New physics-lists for EM standard options 0,1,2,3.
      o RunAction: printout both GetCrossSection() and ComputeCrossSection().
        Reverse order of calls, first ComputeCrossSection(), then
        GetCrossSection().
    + extended/electromagnetic/TestEm1
      o RunAction: replaced 'ProcessCount' with a map.
      o Introduction of G4UIExecutive in main().
    + extended/electromagnetic/TestEm2
      o Added histos rms on profiles (histos numbering is changed)
        and new scripts.
    + extended/electromagnetic/TestEm3
      o Added PhysListEmStandardGS and PhysListEmStandardWVI.
      o Added printout of mean number of gamma, e-, e+.
    + extended/electromagnetic/TestEm7
      o Added extra EM option "ionGasModels", which disables effective ion
        charge and enables use of new models G4BraggIonGasModel, 
        G4BetheBlochIonGasModel.
      o Added extra macro ionGasC12.mac.
      o Fixed setup for histograms through JAIDA.
    + extended/electromagnetic/TestEm8
      o PhysicsList, Em8DetectorConstruction: renamed "VertexDetector" with
        "GasDetector"
      o Added cut per region in TestEm8.in.
    + extended/electromagnetic/TestEm11
    + extended/electromagnetic/TestEm12
      o Added lists PhysListStandardGS, PhysListStandardSS, PhysListStandardWVI.
    + extended/electromagnetic/TestEm13
    + extended/electromagnetic/TestEm14
      o PhysicsList: extend 'ProductionCuts' down to 100 eV.
        Removed non pertinent EmOptions.
      o PhysListEmStandard, Livermore, Penelope: all processes registered as
        discrete (i.e., no continuous Eloss).
      o RunAction: replaced 'ProcessCount' with a map.
      o Cleanup in SteppingAction. Updated README.
    + extended/eventgenerator/HepMC
      o Updated to make use of HepMC-2.06.00.
      o Use QGSP physics-list.
    + extended/eventgenerator/particleGun
      o New example demonstrating three different ways of usage of
        G4ParticleGun, shooting primary particles in different cases.
    + extended/exoticphysics/monopole
      o Substituted QGSP list by QGSP_BERT.
      o Added new classes for monopole G4MonopoleEquation,
        G4MonopoleFieldSetup and G4MonopoleTransportation.
      o Added test on ionisation of heavy highly charged particle; some cleanup.
    + extended/hadronic/Hadr01
      o Cleanup of physics-list according to last update to physics_list
        library.
    + extended/medical/DICOM
      o Added function to read nested items (defined and undefined)
      o Removed dependencies on CONQUEST
      o DICOM handler can now read whichever DICOM file (PT,CT,DR,CR,US,MR... )
    + extended/medical/electronScattering
      o Added PhysListEmStandardWVI.
      o RunAction::PrintFluence(): print unormalized fluence per event.
    + extended/medical/electronScattering2
      o New example for simulation of precision electron scattering benchmark
        with minimal user code.
    + extended/medical/fanoCavity
    + extended/medical/fanoCavity2
      o Introduced PhysListEmStandard_option2 (UrbanMscModel93 without options)
        and PhysListEmStandard_WVI (using WentzelVIModel).
      o Added to run01.mac all physics lists choice.
    + extended/parallel/MPI
      o General code review.
      o mpi_interface: OpenMPI as default MPI library instead of LAM/MPI;
        improved /mpi/status command (added waste of CPU, time measurement);
        introduced /mpi/wait command for waiting until "beamOn" ends;
        improved macro file parser (same parser as G4UIbatch; fixed unexpected
        behavior at the end of batch mode); changed "beamOn" behavior for better
        performance (thread-out for interactive, on-thread for batch).
      o exMPI01: use G4UIcsh instead of G4UItcsh since OpenMPI is now default; 
        use G4eMultipleScattering instead of obsolete G4MultipleScattering.
      o exMPI02: use G4UIcsh instead of G4UItcsh since OpenMPI is now default.
    + examples/extended/persistency/gdml/G01
      o Added G4GenericTrap to the solids.gdml sample.
    + examples/extended/persistency/gdml/G03
      o Corrected selection of G4UItcsh in main(), to fix problem on Windows.
    + examples/novice/N04
      o Replaced QGSP list with QGSP_BERT.


 ----------------------------------------------------------------------------
                             Migration Notes
                             --------------- 

  o Obsolete classes G4MultipleScattering, G4UrbanMscModel and G4UrbanMscModel2
    are removed in this release; multiple-scattering particle-based processes,
    G4eMultipleScattering, G4hMultipleScattering and G4MuMultipleScattering are
    now mandatory; G4UrbanMscModel93 is now introduced.
  o The visualization trajectory modeling command
      "/vis/modeling/trajectories/create/drawByParticleID"
    used to default all trajectories to grey, with the user having to
    explicitly set any other color code. While the user can still set any
    wished color code, the default has been improved to:
    [e-/red; e+/blue; proton/cyan; gamma/green; neutron/yellow;
     pi+,pi-,pi0/magenta; other/grey].
  o In some examples, visualization that used to draw trajectories as simple
    lines now also draws step points. Users who wish to disable step point
    drawing can do so by commenting out a line in the visualization macro.

 ----------------------------------------------------------------------------

                             Technical Notes
                             --------------- 

  o Official platforms:
    + Linux, gcc-4.1.2, gcc-4.3.2.
      Tested on 32 bits architectures and 64 bits architectures
      (Intel or AMD) with the Scientific Linux CERN 5 (SLC5) distribution
      (based on RedHat Linux Enterprise 5).
      Versions of Geant4 have also been compiled successfully on other
      Linux distributions, like Debian, Suse or other RedHat systems.
    + MacOSX 10.6, gcc-4.2.1
    + Windows/XP and CygWin Tools with: Visual C++ 9.0 (Visual Studio 2008)

  o More verified configurations:
    + Linux, gcc-4.5.0
    + Linux, Intel-icc 11.1

  o Geant4 9.4-beta-01 requires the installation of CLHEP-2.0.4.6.

 Please refer to the Geant4 User Documentation:
   http://cern.ch/geant4/support/userdocuments.shtml
 for further information about using Geant4.