Possibility to have different binaries in the same tree introduced
[u/mrichter/AliRoot.git] / PYTHIA / doc / jetset74.lis
CommitLineData
fe4da5cc 1*
2* $Id$
3*
4* $Log$
5* Revision 1.1.1.1 1996/03/08 17:32:28 mclareni
6* jetset74
7*
8*
9* This directory was created from jetset74.car patch update
10
11
12
13 23 August 1995
14
15
16 Updates to
17
18 PYTHIA 5.7 and JETSET 7.4
19 Physics and Manual
20
21 Torbjorn Sjostrand
22 Department of theoretical physics 2
23 University of Lund
24 Solvegatan 14A
25 S-223 62 Lund
26 Sweden
27
28-----------------------------------------------------------------
29
301) Introduction.
31
32Since the PYTHIA/JETSET programs are being updated frequently,
33it is also important to keep the documentation up to date.
34The big manual that appeared as CERN-TH.7112/93 described the
35status as of 15 December 1993. The LaTeX file with this manual
36will be updated, though less frequently than the programs.
37Further, it is not very economical to have to get a new copy
38of a long manual just to see if any interesting new features
39have been added recently. Therefore I have here collected the
40main changes that have taken place since the beginning of 1994.
41The changes are indexed by sub-subversion number and date. This
42file will be updated as regularly as the programs themselves.
43It is not intended to be as complete as the ordinary manual,
44but should be sufficient for the intended purpose.
45
46-----------------------------------------------------------------
47
482) Changes in JETSET 7.4
49
50-----------
51
5200, 13 December 1993: baseline version.
53
54-----------
55
5601, 11 February 1994:
57
58LUZDIS has been changed to protect against an overflow in an exponent
59(harmless physicswise, but enough to crash the program on some
60machines).
61
62-----------
63
6402, 7 April 1994:
65
66A possibility has been introduced into LUSHOW to suppress either hard
67or soft radiation, in the connection of radiation off rapidly decaying
68objects. The algorithm used is not exact, but still gives some
69impression of potential effects. The switch ought to have appeared at
70the end of the current list of shower switches, but because of lack of
71space it appears immediately before.
72
73MSTJ(40) : (D=0) possibility to suppress the branching probability for
74a branching q -> q + g (or q -> q + gamma) of a quark produced in the
75decay of an unstable particle with width Gamma, where this width has to
76be specified by the user in PARJ(89). Can be changed for each new
77LUSHOW call.
78= 0 : no suppression, i.e. the standard parton-shower machinery.
79= 1 : suppress radiation by a factor chi(omega) =
80 Gamma**2 / (Gamma**2 + omega**2), where omega is the energy of the
81 gluon (or photon) in the rest frame of the radiating dipole.
82 Essentially this means that hard radiation with omega > Gamma
83 is removed.
84= 2 : suppress radiation by a factor 1 - chi(omega) =
85 omega**2 / (Gamma**2 + omega**2), where omega is the energy of the
86 gluon (or photon) in the rest frame of the radiating dipole.
87 Essentially this means that soft radiation with omega < Gamma
88 is removed.
89
90PARJ(89) : (D=0. GeV) the width of the unstable particle studied for the
91 MSTJ(40) > 0 options; to be set by the user (separately for each
92 LUSHOW call, if need be).
93
94-----
95
96A generic interface has been included to an external tau decay library.
97This should allow the handling of tau polarization, which is not done
98by JETSET. To use this facility you have to set the switch MSTJ(28),
99include your own interface routine LUTAUD and see to it that the dummy
100routine LUTAUD in JETSET is not linked. The dummy routine is there
101only to avoid unresolved external references when no user-supplied
102interface is linked.
103
104MSTJ(28) : (D=0) call to an external tau decay library.
105= 0 : not done, i.e. the internal LUDECY treatment is used.
106= 1 : done whenever the tau mother particle species can be identified,
107 else the internal LUDECY treatment is used. Normally the mother
108 particle should always be identified, but it is possible for
109 a user to remove event history information or to add extra taus
110 directly to the event record, and then the mother is not known.
111= 2 : always done.
112
113CALL LUTAUD(ITAU,IORIG,KFORIG,NDECAY)
114Purpose: to act as an interface between the generic decay routine
115 LUDECY and a user-supplied tau decay library. The latter library
116 would normally know how to handle polarized taus, given the tau
117 polarization, so one task of the interface routine is to construct
118 the tau polarization/helicity from the information available.
119 Input to the routine is provided in the first three arguments,
120 while the last argument and some event record information have
121 to be set before return.
122ITAU : line number in the event record where the tau is stored.
123 The four-momentum of this tau has first been boosted back to the
124 rest frame of the decaying mother and thereafter rotated to move
125 out along the +z axis. This choice of frame should help the
126 calculation of the helicity configuration. After the LUTAUD call
127 the tau (and its decay products) will be rotated and boosted back.
128 However, seemingly, the event record does not conserve momentum
129 at this intermediate stage.
130IORIG : line number where the mother particle to the tau is stored.
131 Is 0 if the mother is not stored. This does not have to mean the
132 mother is unknown, e.g. in semileptonic B decay the mother is a
133 W+-, and its momentum can be obtained by adding the tau and
134 nu_tau momentum, but there is no line in the event record.
135 When several copies of the mother is stored (e.g. one in the
136 documentation section of the event record and one in the main
137 section), IORIG points to the last. If a branchings like
138 tau -> tau + gamma occurs, the 'grandmother' is given, i.e. the
139 mother of the direct tau before branching.
140KFORIG : flavour code for the mother particle. Is 0 if the mother
141 is unknown. The mother would typically be a resonance such as
142 Z0 (23), W+- (+-24), H0 (25), or H+- (+-37).
143 Often the helicity choice would be clear just by the knowledge
144 of this mother species, e.g. W+- vs. H+-. However, sometimes
145 further complications may exist. For instance, the KF code 23
146 represents a mixture of gamma* and Z0; a knowledge of the mother
147 mass (in P(IORIG,5)) would here be required to make the choice
148 of helicities. Further, a W and Z may either be (predominantly)
149 transverse or longitudinal, depending on the production process
150 under study.
151NDECAY : the number of decay products of the tau; to be given by the
152 user. You must also store the KF flavour codes of those decay
153 products in the positions K(I,2), N+1 <= I <= N+NDECAY, of the
154 event record. The corresponding five-momentum (momentum, energy
155 and mass) should be stored in the associated P(I,J) positions,
156 1 <= J <= 5. The four-momenta are expected to add up to the
157 four-momentum of the tau in position ITAU. You should not change
158 the N value or any of the other K or V values (neither for the
159 tau nor for its decay products) since this is automatically done
160 in LUDECY.
161
162-----
163
164In a few places, a dot has been moved from the end of one line to the
165beginning of the next continuation line, or the other way around,
166to keep together tokens such as .EQ. or .AND., since some debuggers
167may otherwise complain.
168
169A source of (harmless) division by zero in LUSHOW has been removed.
170
171-----------
172
17303, 15 July 1994:
174
175The LUBOEI routine has been changed to avoid an unintentional gap
176in the limits of the very first bin.
177
178Further, leptons and photons which are unrelated to the system
179feeling the Bose-Einstein effects do not have their energies and
180momenta changed in the global rescaling step. (Example: W+W- events,
181where one W decays leptonically; before these lepton momenta could be
182slightly changed, but now not.)
183
184-----
185
186The option LUEDIT(16) (used e.g. from PYEVNT) has been improved with
187a more extensive search for missing daughter pointers.
188
189-----
190
191The KLU(I,16) procedure for finding rank has been rewritten to work
192in the current JETSET version, which it did not before. However, note
193that it will only work for MSTU(16)=2. As a general comment, the
194options 14 - 17 of KLU were written at a time when possible event
195histories were less complex, and can not be guaranteed always to work
196today.
197
198-----------
199
20004, 25 August 1994:
201
202LUSHOW has been corrected, so that if t, l or h quarks (or d* or u*
203quarks masked as l or h) are given with masses that vary from event
204to event (a Breit-Wigner shape, e.g.), the current mass rather than the
205nominal mass is used to define the cut-off scales of parton shower
206evolution.
207
208-----
209
210LULOGO has been modified to take into account that a new PYTHIA/JETSET
211description has been published in
212T. Sjostrand, Computer Phys. Commun. 82 (1994) 74
213and is from now on the standard reference to these two programs.
214
215-----------
216
21705, 27 January 1995:
218
219LUCELL has been corrected, in that in the option with smearing of
220energy rather than transverse energy, the conversion factor between
221the two was applied in the wrong direction.
222
223-----
224
225LUSHOW has been corrected in one place where the PMTH array was
226addressed with the wrong order of the indices. This affected quark
227mass corrections in the matching to the three-jet matrix elements.
228
229-----
230
231An additional check has been included in LUBOEI that there are at
232least two particles involved in the Bose-Einstein effects. (No
233problem except in some bizarre situations.)
234
235-----------
236
23706, 20 February 1995:
238
239A new option has been added for the behaviour of the running
240alpha_em(Q2) in ULALEM. This is not added as a true physics scenario,
241but only to produce results with a given, fixed value for the hard
242events, while still keeping the conventional value in the Q2=0 limit.
243MSTU(101) = 2 : if Q2 is less than PARU(104) then alpha_em is
244 assigned the value PARU(101) (=1/137), while for Q2 above
245 PARU(104) the fixed value PARU(103) (=1/128.8) is used.
246PARU(103) : (D=0.007764=1/128.8) alpha_em used for hard processes
247 in the option MSTU(101)=2.
248PARU(104) : (D=1 GeV^2) dividing line for 'low' and 'high'
249 Q2 values in the MSTU(101)=2 option of ULALEM.
250
251Additionally, the G_F constant has been added to the parameter list.
252PARU(105) : (D=1.16639E-5 GeV^-2) G_F, the Fermi constant of weak
253 interactions.
254
255-----
256
257The LULOGO routine has been updated to reflect my change of
258affiliation.
259
260-----------
261
26207, 21 June 1995:
263
264Header and LULOGO have been updated with respect to phone number
265and WWW access.
266
267-----
268
269The PHEP and VHEP variables in the /HEPEVT/ common block are now
270assumed to be in DOUBLE PRECISION, in accord with the proposed
271LEP 2 workshop addendum to the standard.
272
273-----
274
275In LUTEST a missing decimal point on the energy check has been
276reinstated (0001 -> 0.0001).
277
278-----
279
280In LUINDF the expression PR/(Z*W) has been protected against vanishing
281denominator.
282
283-----------
284
28508, 23 August 1995:
286
287Check against division by zero in LUSHOW.
288
289-----------------------------------------------------------------
290
2913) Changes in PYTHIA 5.7
292
293-----------
294
29500, 13 December 1993: baseline version.
296
297-----------
298
29901, 27 January 1994:
300
301The machinery to handle gamma-gamma interactions is expanded.
302In particular, several new options have been added to MSTP(14).
303The updated description of this variable reads as follows.
304
305MSTP(14) : (D=0) structure of incoming photon beam or target
306 (does not affect photon inside electron, only photons appearing
307 as argument in the PYINIT call).
308 = 0 : a photon is assumed to be point-like (a direct photon),
309 i.e. can only interact in processes which explicitly contain
310 the incoming photon, such as f_i gamma -> f_i g for gamma-p
311 interactions. In gamma-gamma interactions both photons are
312 direct, i.e the main process is gamma gamma -> f_i fbar_i.
313 = 1 : a photon is assumed to be resolved, i.e. can only interact
314 through its constituent quarks and gluons, giving either high-pT
315 parton-parton scatterings or low-pT events. Hard processes are
316 calculated with the use of the full photon parton distributions.
317 In gamma-gamma interactions both photons are resolved.
318 = 2 : a photon is assumed resolved, but only the VMD piece is
319 included in the parton distributions, which therefore mainly
320 are scaled-down versions of the rho0/pi0 ones. Both high-pT
321 parton-parton scatterings and low-pT events are allowed. In
322 gamma-gamma interactions both photons are VMD-like.
323 = 3 : a photon is assumed resolved, but only the anomalous piece of
324 the photon parton distributions is included. Only high-pT
325 parton-parton scatterings are allowed. In gamma-gamma
326 interactions both photons are anomalous.
327 = 4 : in gamma-gamma interactions one photon is direct and the other
328 resolved. A typical process is thus f_i gamma -> f_i g. Hard
329 processes are calculated with the use of the full photon
330 parton distributions for the resolved photon. Both possibilities
331 of which photon is direct are included, in event topologies and
332 in cross sections. This option cannot be used in configurations
333 with only one incoming photon.
334 = 5 : in gamma-gamma interactions one photon is direct and the other
335 VMD-like. Both possibilities of which photon is direct are
336 included, in event topologies and in cross sections. This option
337 cannot be used in configurations with only one incoming photon.
338 = 6 : in gamma-gamma interactions one photon is direct and the other
339 anomalous. Both possibilities of which photon is direct are
340 included, in event topologies and in cross sections. This option
341 cannot be used in configurations with only one incoming photon.
342 = 7 : in gamma-gamma interactions one photon is VMD-like and the other
343 anomalous. Only high-pT parton-parton scatterings are allowed.
344 Both possibilities of which photon is VMD-like are included,
345 in event topologies and in cross sections. This option cannot be
346 used in configurations with only one incoming photon.
347 Note: a complete description requires separate runs for the components
348 above, i.e. it is not possible to mix them in a single run. Our
349 best understanding of gamma-p interactions [Sch93,Sch93a] is to
350 have three separate components, 0 + 2 + 3. A simpler alternative
351 is based on two only, 0 + 1. Our best understanding of gamma-gamma
352 interactions [in preparation] requires six separate components,
353 0 + 2 + 3 + 5 + 6 + 7. A simpler alternative is based on three
354 only, 0 + 1 + 4.
355
356In addition, one new option has been introduced and a few internal
357variables modified.
358
359MSTP(59) : (D=0) possibility to modify the Q2 scale used in the
360 anomalous parton distributions of the photon, as used in the
361 options MSTP(14) = 3, 6 and 7.
362 = 0 : no change of Q2 scale compared to what is normally used.
363 = 1 : the input Q2 scaled is divided by PARP(59)**2 to define
364 the Q2 scale used as argument for the anomalous parton
365 distributions.
366
367PARP(59) : (D=1.) rescaling factor used for the Q2 argument of the
368 anomalous parton distributions of the photon, see MSTP(59).
369
370MINT(105) : is MINT(103) or MINT(104), depending on which side
371 of the event currently is being studied.
372
373MINT(107), MINT(108) : if either or both of the two incoming particles
374 is a photon, then the respective value gives the nature assumed for
375 that photon. The code follows the one used for MSTP(14):
376 = 0 : direct photon.
377 = 1 : resolved photon.
378 = 2 : VMD-like photon.
379 = 3 : anomalous photon.
380
381MINT(109) : is either MINT(107) or MINT(108), depending on which side
382 of the event currently is being studied.
383
384VINT(282) : no longer used.
385
386VINT(283), VINT(284): virtuality scale at which an anomalous photon
387 on the beam or target side of the event is being resolved. More
388 precisely, it gives the p_T^2 pf the gamma -> q qbar vertex.
389
390-----
391
392A number of bugs have also been corrected:
393* Jet + low-pT event generation could give incorrect cross section
394 information with PYSTAT(1) at low energies. The event generation
395 itself is correct. (The error was introduced when variable energies
396 became allowed.)
397* Introduce rejection of top events where top mass (in the tails of the
398 Breit-Wigner distribution) is too low to allow decays t -> W + b.
399* Plus a few minor bugs, probably harmless.
400
401-----------
402
40302, 13 February 1994:
404
405The interface to PDFLIB has been modified to reflect that 'TMAS' should
406no longer be set except in first PDFSET call. (Else a huge amount of
407irrelevant warning messages are generated by PDFLIB.)
408
409-----
410
411The STOP statement in a few dummy routines has been modifed to avoid
412irrelevant compilation warning messages on IBM mainframes.
413
414-----
415
416A few labels have been renumbered.
417
418-----------
419
42003, 22 February 1994:
421
422Removal of a bug in PYRESD, which could give (under some specific
423conditions) errors in the colour flow.
424
425-----------
426
42704, 7 April 1994:
428
429Process 11 has been corrected, for the part that concerns anomalous
430couplings (contact interactions) in the q + q' -> q + q' process.
431The error was present in the expression for u + dbar -> u + dbar
432and obvious permutations, while u + d -> u + d, u + ubar -> u + ubar
433and the others were correct. Thanks to J.-J. Dugne, M. Perrottet and
434K. Lane for communications on this point.
435
436-----
437
438The option MSTP(23)=1 for post-facto (x,Q^2) conservation in deep
439inelastic scattering can give infinite loops when applied to process
44083, in particular if one asks for the production of a top. (Remember
441that the standard DIS kinematics only is defined for massless quarks.)
442Therefore the switch MSTP(23) has been modifed as follows:
443MSTP(23) : (D=1) (x, Q^2) correction level in DIS.
444 = 0 : no correction procedure applied.
445 = 1 : correction applied for process 10, but not for process 83.
446 = 2 : correction applied both for process 10 and 83. This latter
447 option could still work fine for charm and bottom, if the
448 energy is sufficient.
449
450-----
451
452PYRESD is modified to ensure isotropic angular distributions in the
453decays of the top or a fourth generation particle, i.e. in t -> b + W+.
454This may not be the correct distribution but, unless explicit knowledge
455exists for a given process, this should always be the default.
456
457-----
458
459In processes 16, 20, 31 and 36 the W propagator has been modified to
460include s-dependent widths in the Breit-Wigner shape. The most notable
461effect is a suppression of the low-mass tail of the W mass spectrum.
462
463-----
464
465When PDFLIB is used, PDFSET is now only called whenever a different
466structure function is requested. For pp events therefore only one call
467is made, while gamma-p interactions still involves a call to PDFSET
468for each STRUCTM one, since gamma and p structure functions have to be
469called alternatingly. To this end, MINT(93) is reset to
4701000000 * Nptype + 1000 * Ngroup + Nset after each PDFSET call.
471
472-----
473
474In a few places, a dot has been moved from the end of one line to the
475beginning of the next continuation line, or the other way around,
476to keep together tokens such as .EQ. or .AND., since some debuggers
477may otherwise complain. Also some other purely cosmetics changes
478for the same reason.
479
480A number of minor errors have been corrected.
481
482-----------
483
48405, 15 July 1994:
485
486A new option has been introduced, MSTP(14)=10, whereby it is possible
487to obtain a mixture of the various allowed photon components. For
488gamma-hadron collisions, this means a mixture of VMD, direct and
489anomalous events, for gamma-gamma collisions a mixture of VMD*VMD,
490VMD*direct, VMD*anomalous, direct*direct, direct*anomalous and
491anomalous*anomalous. The mixture is properly given according to
492the relative cross sections.
493
494Note that this introduces a completely new layer of administration in
495PYTHIA. For instance, a subprocess such as q + g -> q + g is allowed
496in the VMD*VMD, VMD*anomalous and anomalous*anomalous classes, but
497appear with different sets of parton distributions and with different
498pT cut-offs. In order to handle this, various information is initialized
499separately for each event class, and subsequently saved and restored
500as the generation switches back and forth between the event classes.
501This introduces some limitations on what you may and may not do.
502
503First of all, the MSTP(14) switch is only applicable for incoming photon
504beams, i.e. when 'gamma' is the argument in the PYINIT call. A
505convolution with the bremsstrahlung photon spectrum in an electron beam
506may come one day, but not in the immediate future.
507
508Secondly, the machinery has only been set up to generate standard
509QCD physics, specifically either 'minimum bias' one or high-pT jets.
510For minimum bias, you are not allowed to use the CKIN variables at all.
511This is not a major limitation, since it is in the spirit of minimum
512bias physics not to impose any contraints on allowed jet production.
513(If you still do, these cuts will be ineffective for the VMD processes
514but take effect for the other ones, giving inconsistencies.) Further,
515some variables are internally recalculated and reset: CKIN(1), CKIN(3),
516CKIN(5), CKIN(6), MSTP(57), MSTP(85), PARP(2), PARP(81), PARP(82),
517PARU(115) and MDME(22,J). These can not be modified without changing
518PYINPR and recompiling the program. The minimum bias physics option
519is obtained by default; by switching from MSEL=1 to MSEL=2 also the
520elastic and diffractive components of the VMD part are included.
521High-pT jet production is obtained by setting the CKIN(3) cut-off
522larger than the (energy-dependent) cut-off scales for the VMD, direct
523and anomalous components; typically this means at least 3 GeV. For
524lower input CKIN(3) the program will automatically switch back to
525minimum bias physics.
526
527Finally, pileup events are not at all allowed.
528
529Here is a survey of common block variables affected:
530
531MSTP(14) (D=0) strucure of incoming photon beam or target;
532 see description above for PYTHIA 5.701.
533 = 10 : new option where the VMD, direct and anomalous components
534 are automatically mixed, as described above. Works equally well
535 for gamma-p and gamma-gamma.
536
537MSTI(9) : event class used in current event.
538 = 1 : VMD (for gamma-p) or VMD*VMD (for gamma-gamma).
539 = 2 : direct (for gamma-p) or VMD*direct (for gamma-gamma).
540 = 3 : anomalous (for gamma-p) or VMD*anomalous (for gamma-gamma).
541 = 4 : direct*direct (for gamma-gamma).
542 = 5 : direct*anomalous (for gamma-gamma).
543 = 6 : anomalous*anomalous (for gamma-gamma).
544
545MINT(121) : number of separate event classes to initialize and mix.
546 = 1 : the normal value.
547 = 3 : for a gamma-hadron interaction when MSTP(14)=10.
548 = 6 : for a gamma-gamma interaction when MSTP(14)=10.
549
550MINT(122) : event class used in current event. Code as explained for
551 MSTI(9).
552
553MINT(123) : event class used in the current event, with the same list
554 of possibilities as MSTP(14), except that MSTP(14) = 1, 4 or 10
555 do not appear.
556
557VINT(285) : the CKIN(3) value provided by the user at initialization;
558 subsequently CKIN(3) may be overwritten (for MSTP(14)=10) but
559 VINT(285) stays.
560
561In addition, the structure of the initialization has been partly
562reorganized. The routine PYEVKI has been removed, new routines
563PYINBM, PYINPR and PYSAVE created, and some material has been moved
564to or from PYINIT, PYINRE and PYINKI.
565
566SUBROUTINE PYINBM : to read in and identify beam and target particles
567 and frame as given in the PYINIT call (used to be done in PYINKI).
568
569SUBROUTINE PYINKI(MODKI) : to set up event kinematics, either at
570 initialization (MODKI=0) or for each separate event when varying
571 kinematics (MODKI=1). (The latter task used to be done in PYEVKI.)
572
573SUBROUTINE PYINPR : to set up the partonic subprocesses selected with
574 MSEL and, for gamma-p and gamma-gamma, MSTP(14).
575
576SUBROUTINE PYSAVE : saves and restores parameters and cross section
577 values between the 3 gamma-p and the 6 gamma-gamma alternatives
578 of MSTP(14)=10. Also makes a random choice for each new event
579 between the allowed alternatives.
580
581Among other changes, note that PYSTAT(1) now has been extended so
582that the subdivision into the various gamma-p and gamma-gamma classes
583is shown.
584
585-----
586
587Further changes of particular relevance for gamma-p and gamma-gamma,
588but independent of the major revisions above:
589
590MSTP(59) and PARP(59) have been removed. Instead the following options
591are available:
592
593MSTP(15) : (D=5) possibility to modify the nature of the anomalous
594photon component, in particular with respect to the scale choices and
595cut-offs of hard processes.
596 = 0 : none, i.e. the same treatment as for the VMD component.
597 = 1 : evaluate the anomalous structure functions at a scale
598 Q2/PARP(17)^2.
599 = 2 : as =1, but instead of PARP(17) use PARP(81)/PARP(15) or
600 PARP(82)/PARP(15), depending on MSTP(82) value.
601 = 3 : evaluate anomalous structure function as
602 f^(anom)(x, Q2, p_0^2) - f^(anom)(x, Q2, r^2*Q2)
603 with r = PARP(17).
604 = 4 : as =3, but instead of PARP(17) use PARP(81)/PARP(15) or
605 PARP(82)/PARP(15), depending on MSTP(82) value.
606 = 5 : use larger pTmin for anomalous component than for VMD one,
607 but otherwise no difference.
608
609PARP(17) : (D=1) rescaling factor used as described for MSTP(15).
610
611MSTP(51) : new option added.
612 = 11 : the GRV p LO parametrization.
613
614MSTP(53) : new option added.
615 = 3 : the GRV pi LO parametrization.
616
617MSTP(56) : new option added.
618 = 3 : when the anomalous photon structure function is requested,
619 the homogeneous solution is provided, evolved from a starting
620 value PARP(15) to the requested Q scale. The homogeneous
621 solution is normalized so that the net momentum is unity,
622 i.e. any factors of alpha_em/2pi and charge have been left out.
623 The flavour of the original q is given in MSTP(55) (1, 2, 3, 4
624 or 5 for d, u, s, c or b); the value 0 gives a mixture
625 according to squared charge, with the exception that c and b
626 are only allowed above the respective mass threshold (Q > m_q).
627 The four-flavour Lambda value is assumed given in PARP(1);
628 it is automatically recalculated for 3 or 5 flavours at
629 thresholds. This option is not intended for standard event
630 generation, but is useful for some theoretical studies.
631
632-----
633
634Option MSTP(92)=5 for beam remnant treatment erroneously missed some
635statements which now have been inserted.
636
637Further, new options have been added for the splitting of momentum
638between two beam remnants. MSTP(92) keeps its current role for the
639production of diquark or quark jets. However, for the splitting into
640a hadron plus a quark/diquark jet, MSTP(94) should now be used.
641
642MSTP(94) : (D=2) (C) energy partitioning in hadron or resolved photon
643remnant, when this is split into a hadron plus a remainder-jet. The
644energy fraction chi is taken by one of the two objects, with
645conventions as described below or for PARP(95) and PARP(97).
646 = 1 : 1 for meson or resolved photon, 2(1-chi) for baryon, i.e.
647 simple counting rules.
648 = 2 : (k+1)*(1-chi)**k, with k as given in PARP(95) or PARP(97).
649 = 3 : the chi of the hadron is selected according to the normal
650 fragmentation function used for the hadron in jet fragmentation,
651 see MSTJ(11). The possibility of a changed fragmentation
652 function shape in diquark fragmentation (see PARJ(45)) is not
653 included.
654 = 4 : as =3, but the shape is changed as allowed in diquark
655 fragmentation (see PARJ(45)); this change is here also allowed
656 for meson production.
657
658-----
659
660In PYDIFF the recoiling gluon energy is calculated in a numerically more
661stable fashion.
662
663-----
664
665A counter has been added to PYSSPA to avoid infinite loops in the
666angular ordering constraint due to interference with the final state
667colour charges.
668
669-----------
670
67106, 25 August 1994:
672
673New processes 167 and 168 have been added for the contact interaction
674production of d* or u* excited quarks
675167 q q' -> q" d*
676168 q q' -> q" u*
677where the different allowed quark and antiquark combinations are given
678according to eqs. (15) - (19) in U. Baur, M. Spira and P.M. Zerwas,
679Phys. Rev. D42 (1990) 815. The d* and u* are defined in the same
680way as for processes 147 and 148. Thus one needs to put MSTP(6)=1 to
681use l (7) and h (8) for representing the d* and u*. The couplings of
682the allowed decay channels are given by PARU(157) - PARU(159), and
683the Lambda scale parameter by PARU(155).
684
685At the same time, some minor changes has been introduced in the code
686for processes 147 and 148, for uniformity.
687
688-----
689
690Option MSTP(57)=3 now also allows a dampening of pi+- parton
691distributions.
692
693-----
694
695A few minor errors have been corrected.
696
697-----------
698
69907, 20 October 1994:
700
701A major bug discovered in processes 121 and 122 (and thus also affecting
702181, 182, 186 and 187), g g or q qbar -> Q Qbar H: the kinematics was
703incorrectly handed on to the Kunszt matrix elements. This affected the
704default option Q = t, but effects were especially dramatic when the
705alternative Q = b was used.
706
707The choice of appropriate Q2 scale for structure functions introduces
708a further uncertainty in cross sections for the processes above. So long
709as only t quarks are considered, the t mass is a reasonable choice, but
710for the Q = b alternative this is presumably too low. Therefore new
711options have been introduced as below, with the default behaviour
712changed (the old one is obtainable with MSTP(39)=1).
713
714MSTP(39) : (D=2) choice of Q2 scale for structure functions and initial
715state parton showers in processes g g or q qbar -> Q Qbar H.
716 = 1 : m_Q**2.
717 = 2 : max(mT_Q**2 , mT_Qbar**2) =
718 m_Q**2 + max(pT_Q**2 , pT_Qbar**2).
719 = 3 : m_H**2.
720 = 4 : shat = (p_H + p_Q + p_Qbar)**2.
721
722-----
723
724Another important bug corrected in the calculation of the reduction of
725t+tbar cross section when decay modes are forced. This occured when both
726t and tbar produced a W, and W+ and W- decay modes were set differently.
727
728-----------
729
73008, 25 October 1994:
731
732A few further places changed to make processes 181, 182, 186 and 187
733work (see version 5.707 above).
734
735-----------
736
73709, 26 October 1994:
738
739The matrix element for f + fbar -> W+ + W- has been replaced, using the
740formulae of
741D. Bardin, M. Bilenky, D. Lehner, A. Olchevski and T. Riemann,
742CERN-TH.7295/94,
743but with the dependence on the t-hat variable not integrated out
744(D. Bardin, private communication).
745This avoids some problems encountered in the old expressions when
746one or both W's were far off the mass shell.
747
748-----
749
750Change in calls to PDFLIB, so that the input Q is always at least the
751QMIN of the respective set.
752
753-----
754
755Extra protection against infinite loops in PYSSPA.
756
757-----------
758
75910, 27 January 1995:
760
761The dimensions of the HGZ array in PYRESD has been expanded to avoid
762accidental writing outside the bounds.
763
764-----
765
766VINT(41) - VINT(66) are saved and restored in PYSCAT, for use in low-pT
767events, when beam remnant treatment has failed (with nonzero MINT(57)).
768
769-----
770
771The routine PYSTGH has been replaced by the routine PYSTHG. This
772contains an improved parametrization of the homogeneous evolution
773of an anomalous photon from some given initial scale. The argument
774NF of the PYSTGH routine has been removed; now Lambda is always
775automatically converted to the relevant NF-flavour value from its
7764-flavour one, at flavour thresholds.
777
778-----------
779
78011, 20 February 1995:
781
782New possibilities have been added to switch between electroweak
783couplings being expressed in terms of a running alpha_em(Q2) or
784in terms of a fixed Fermi constant G_F. This affects both decay widths
785and process cross sections, in the routines PYINRE, PYRESD, PYWIDT and
786PYSIGH. There are three main options, with default agreeing with the
787old standard.
788
789MSTP(8) : (D=0) choice of electroweak parameters to use in decay
790 widths of resonances (W, Z, H, ...) and cross sections (production
791 of W's, Z's, H's, ...).
792 = 0 : everything is expressed in terms of a running alpha_em(Q2)
793 and a fixed sin^2(theta_W), i.e. G_F is nowhere used.
794 = 1 : a replacement is made according to
795 alpha_em(Q2) -> sqrt(2) G_F m_W^2 sin^2(theta_W) / pi
796 in all widths and cross sections. If G_F and m_Z are considered
797 as given, this means that sin^2(theta_W) and m_W are the only
798 free electroweak parameter.
799 = 2 : a replacement is made as for =1, but additionally
800 sin^2(theta_W) is constrained by the relation
801 sin^2(theta_W) = 1 - m_W^2/m_Z^2
802 This means that m_W remains as a free parameter, but that the
803 sin^(theta_W) value in PARU(102) is never used, EXCEPT in
804 the vector couplings in the combination
805 v = a - 4 sin^2(theta_W) e.
806 This degree of freedom enters e.g. for forward-backward
807 asymmetries in Z^0 decays.
808Note : This option does not affect the emission of real photons in the
809 initial and final state, where alpha_em is always used. However,
810 it does affect also purely electromagnetic hard processes, such as
811 q + qbar -> gamma + gamma.
812
813-----
814
815The option MSTP(37)=1, with running quark masses in couplings to Higgs
816bosons, only works when alpha_s is allowed to run (so one can define
817a Lambda value). Therefore a check has been introduced in PYWIDT and
818PYSIGH that the option MSTP(37)=1 is only executed if additionally
819MSTP(2) is 1 or bigger.
820
821-----
822
823Some non-physics changes have been made in the RKBBV and STRUCTM codes
824so as to avoid some (in principle harmless) compiler warnings.
825
826-----------
827
82812, 15 March 1995:
829
830A serious error has been corrected in the MSTP(173)=1 option, i.e. when
831the program is run with user-defined weights that should compensate for
832a biased choice of variable beam energies. This both affected the
833relative admixture of low- and high-pT events and the total cross
834section obtained by Monte Carlo integration. (PYRAND changed.)
835
836In order to improve the flexibility and efficiency of the variable-energy
837option, the user should now set PARP(174) before the PYINIT call, and
838thereafter not change it. This allows PARP(173) weights of arbitrary
839size. (PYRAND and PYMAXI changed.)
840PARP(174) : (D=1.) maximum event weight that will be encountered in
841 PARP(173) during the course of a run with MSTP(173)=1; to be used
842 to optimize the efficiency of the event generation. It is always
843 allowed to use a larger bound than the true one, but with a
844 corresponding loss in efficiency.
845
846MSTI(5) (and MINT(5)) are now changed so they count the number of
847successfully generated events, rather than the number of tries made.
848This change only affects runs with variable energies, MSTP(171)=1 and
849MSTP(172)=2, where MSTI(61)=1 signals that a user-provided energy
850has been rejected in the weighting. This change also affects PARI(2),
851which becomes the cross section per fully generated event. (PYEVNT
852changed.)
853
854-----
855
856The option MSTP(14)=10 has now been extended so that it also works for
857deep inelastic sacattering of an electron off a (real) photon, i.e.
858process ISUB = 10. What is obtained is a mixture of the photon acting
859as a vector meson and it acting as an anomalous state. This should
860therefore be the sum of what can be obtained with MSTP(14)=2 and =3.
861It is distinct from MSTP(14)=1 in that different sets are used for
862the parton distributions - in MSTP(14)=1 all the contributions to the
863photon distributions are lumped together, while they are split in
864VMD and anomalous parts for MSTP(14)=10. Also the beam remnant treatment
865is different, with a simple Gaussian distribution (at least by default)
866for MSTP(14)=1 and the VMD part of MSTP(14)=10, but a powerlike
867distribution d(kT^2)/kT^2 between PARP(15) and Q for the anomalous
868part of MSTP(14)=10. (PYINIT, PYINPR and PYSTAT changed.)
869
870To access this option for e and gamma as incoming beams, it is only
871necessary to set MSTP(14)=10 and keep MSEL at its default value.
872Unlike the corresponding option for gamma-p and gamma-gamma, no
873cuts are overwritten, i.e. it is still the responsability of the user
874to set these appropriately. Those especially appropriate for DIS usage
875are CKIN(21)-CKIN(22) or CKIN(23)-CKIN(24) for the x range (former or
876latter depending on which side is the incoming real photon),
877and CKIN(35)-CKIN(36) for the Q2 range. A further new option has been
878added (in PYKLIM):
879CKIN(39), CKIN(40) : (D=4., -1. GeV^2) the W2 range allowed in DIS
880 processes, i.e. subprocess number 10. If CKIN(40) < 0., the upper
881 limit is inactive. Here W2 is defined in terms of W2 = Q2 * (1-x)/x.
882 This formula is not quite correct, in that (i) it neglects the
883 target mass (for a proton), and (ii) it neglects initial-state
884 photon radiation off the incoming electron. It should be good
885 enough for loose cuts, however.
886
887A warning about the usage of PDFLIB for photons. So long as MSTP(14)=1,
888i.e. the photon is not split up, PDFLIB is accessed by MSTP(56)=2 and
889MSTP(55) = parton distribution set, as described in the manual.
890However, when the VMD and anomalous pieces are split, the VMD part
891is based on a rescaling of pion distributions by VMD factors
892(except for the SaS sets, that already come with a separate VMD piece).
893Therefore, to access PDFLIB for MSTP(14)=10, it is not correct to set
894MSTP(56)=2 and a photon distribution in MSTP(55). Instead, one should
895put MSTP(56)=2, MSTP(54)=2 and a pion distribution code in MSTP(53),
896while MSTP(55) has no function. The anomalous part is still based on
897the SaS parametrization, with PARP(15) as main free parameter.
898
899-----
900
901A change has been made in PYREMN to reduce the possibility of infinite
902loops.
903
904-----------
905
90613, 22 March 1995:
907
908The SaS parton distributions of the photons are now available.
909For details on these sets, see
910G.A. Schuler and T. Sjostrand,
911"Low- and high-mass components of the photon distribution functions",
912CERN-TH/95-62 and LU TP 95-6.
913There are four new sets. These differ in that two use a Q0=0.6 GeV and
914two a Q0=2 GeV, and in that two use the DIS and two the MSbar conventions
915for the dominant non-leading contributions. (However, the fits are
916formally still leading-order, in that not all next-to-leading
917contributions have been included.) New default is the SaS 1D set.
918Furthermore, for the definition of F_2^gamma, additional terms appear
919that do not form part of the parton distributions itself. To partly
920take this into account, an additional doubling of the possibilities
921has been included. These possibilites can be accesed with MSTP(55):
922
923MSTP(55) : (D=5) choice of parton-distribution set of the photon;
924 see also MSTP(56).
925 = 1 : Drees-Grassie.
926 = 5 : SaS 1D (in DIS scheme, with Q0=0.6 GeV).
927 = 6 : SaS 1M (in MSbar scheme, with Q0=0.6 GeV).
928 = 7 : SaS 2D (in DIS scheme, with Q0=2 GeV).
929 = 8 : SaS 2M (in MSbar scheme, with Q0=2 GeV).
930 = 9 : SaS 1D (in DIS scheme, with Q0=0.6 GeV).
931 = 10 : SaS 1M (in MSbar scheme, with Q0=0.6 GeV).
932 = 11 : SaS 2D (in DIS scheme, with Q0=2 GeV).
933 = 12 : SaS 2M (in MSbar scheme, with Q0=2 GeV).
934 Note 1 : sets 5 - 8 use the parton distributions of the respective
935 set, and nothing else. These are appropriate for most
936 applications, e.g. jet production in gamma-p and gamma-gamma
937 collisions. Sets 9 - 12 instead are appropriate for
938 gamma*-gamma processes, i.e. DIS scattering on a photon,
939 as measured in F_2^gamma. Here the anomalous contribution
940 for c and b quarks are handled by the Bethe-Heitler formulae,
941 and the direct term is artificially lumped with the anomalous
942 one, so that the event simulation more closely agrees with what
943 will be experimentally observed in these processes. The agreement
944 with the F_2^gamma parametrization is still not perfect, e.g.
945 in the treatment of heavy flavours close to threshold.
946 Note 2 : Sets 5 - 12 contain both VMD pieces and anomalous pieces,
947 separately parametrized. Therefore the respective piece is
948 automatically called, whatever MSTP(14) value is used to select
949 only a part of allowed photon interactions. For other sets
950 (set 1 above or PDFLIB sets), usually there is no corresponding
951 subdivision. Then an option like MSTP(14)=2 (VMD part of photon
952 only) is based on a rescaling of the pion distributions, while
953 MSTP(14)=3 gives the SaS anomalous parametrization.
954 Note 3 : Formally speaking, the k0 (or p0) cut-off in PARP(15)
955 need not be set in any relation to the Q0 cut-off scales used by
956 the various parametrizations. Indeed, due to the familiar scale
957 choice ambiguity problem, there could well be some offset between
958 the two. However, unless you know what you are doing, it is
959 strongly recommended that you let the two agree, i.e. set
960 PARP(15)=0.6 for the SaS 1 sets and =2 for the SaS 2 sets.
961
962PARP(15) : (D=0.6 GeV) default value changed for k0 cut-off for
963 separation between direct, VMD and anomalous photons; see Note 3
964 for MSTP(55) above.
965
966The generic routine PYSTFU has been rewritten to handle the interfacing.
967The old routines PYSTAG, PYSTGS, PYDILN and PYSTHG have been removed.
968Instead the routines of the SaSgam library have been inserted. In order
969to avoid any clashes, the routines SAS*** have been renamed PYG***.
970Thus new routines are PYGGAM, PYGVMD, PYGANO, PYGBEH and PYGDIR.
971The common block SASCOM is renamed PYINT8. If you want to use the
972parton distributions for standalone purposes, you are encouraged to
973use the original SaSgam routines rather than going the way via the
974Pythia adaptations.
975
976 COMMON/PYINT8/XPVMD(-6:6),XPANL(-6:6),XPANH(-6:6),XPBEH(-6:6),
977 &XPDIR(-6:6)
978Purpose: to store the various components of the parton distributions
979 when the PYGGAM routine is called.
980XPVMD(KFL) : gives distributions of the VMD part (rho, omega and
981 phi).
982XPANL(KFL) : gives distributions of the anomalous part of light quarks
983 (d, u and s).
984XPANH(KFL) : gives distributions of the anomalous part of heavy quarks
985 (c and b).
986XPBEH(KFL) : gives Bethe-Heitler distributions of heavy quarks (c and b).
987 This provides an alternative to XPANH, i.e. both should not be used
988 at the same time.
989XPDIR(KFL) : gives direct correction to the production of light quarks
990 (d, u and s). This term is nonvanishing only in the MSbar scheme,
991 and is applicable for F_2^gamma rather than for the parton
992 distributions themselves.
993
994-----
995
996PYDOCU has been corrected so that PARI(2) refers to the full cross
997section for gamma-p and gamma-gamma processes, rather than that of
998the latest subprocess considered.
999
1000An additional check has been inserted into PYREMN.
1001
1002-----------
1003
100414, 23 March 1995:
1005
1006Some minor modifications to PYSTFU and PYGGAM in the wake of the
1007changes of the previous version.
1008
1009-----------
1010
101115, 24 April 1995:
1012
1013An unfortunate choice of default values has been corrected:
1014the old MSTP(3)=2 value implied that Lambda_QCD was entirely based on
1015the Lambda value of the proton structure function; also e.g. for e+e-
1016annihilation events. Thus the Lambda in PARJ(81) was overwritten,
1017i.e. did not keep the value required by standard phenomenology, which
1018typically gave too narrow jets. (While switching to MSTP(3)=1 it worked
1019fine.) In the modified option MSTP(3)=2 this has been corrected, to
1020better agree with user expectations. Change affects PYINIT and
1021PYRESD. (See further version 16 for additional changes.)
1022
1023MSTP(3) : (D=2) choice of Lambda_QCD values.
1024 = 1 : separate for hard scattering, initial showers and final
1025 showers, as before. Additionally separate for resonance
1026 decays, given in PARP(3).
1027 = 2 : most Lambda values are set to that of the proton structure
1028 function used, except for the Lambda used in the decay of
1029 a resonance (as treated in PYRESD). There the PARP(3) value
1030 is used, with default as in e+e-.
1031 = 3 : all Lambda values are set to that of the proton structure
1032 function used, as was the case for =2 before.
1033
1034PARP(3) : (D=0.29 GeV) the Lambda value used in timelike parton showers
1035 in the decay of a resonance (in PYRESD).
1036
1037-----
1038
1039The form for PTMANO, the pTmin for anomalous processes, as used in
1040PYINPR when processes are mixed for gamma-p or gamma-gamma events,
1041has been updated to match (as well as can be expected) the SaS 1D
1042photon distributions.
1043
1044-----------
1045
104616, 30 June 1995:
1047
1048The strategy for the changes to MSTP(3) in version 15 above have been
1049modified for better transparency. The parameter PARP(3) has been removed,
1050and instead PARP(72) has been introduced. Now PARJ(81) is used for
1051resonance decays (including e.g. Z0 decay, from which it is determined),
1052and PARP(72) for other timelike showers. PARJ(81) is not overwritten
1053for MSTP(3) = 2, but only for = 3. Changes affect PYINIT, PYEVNT and
1054PYRESD.
1055
1056PARP(72) : (D=0.25 GeV) the Lambda value used in timelike parton showers
1057 except in the decay of a resonance.
1058
1059-----
1060
1061A new multiplicative factor has been introduced for the hard scattering
1062in PYSIGH.
1063
1064PARP(34) : (D=1.) the Q2 scale defined by MSTP(32) is multiplied by PARP(34)
1065 when it is used as argument for structure functions and alpha_s at the
1066 hard interaction. It does not affect alpha_s when MSTP(33)=3, nor
1067 does it change the Q2 argument of parton showers.
1068
1069-----
1070
1071PYREMN has been corrected for occasional too large boost factors.
1072
1073An error in PYSIGH for process 148 has been corrected.
1074
1075The MSTP(62)=1 option of PYSSPA is modified to avoid division by zero.
1076
1077Header has been updated with WWW-information.
1078
1079-----------
1080
108117, 23 August 1995:
1082
1083MIN1, MIN2, MAX1, MAX2, MINA and MAXA in PYSIGH have had an extra M
1084prefixed to avoid confusion with Fortran functions.
1085
1086Protect against MDCY(0,1) being accessed in PYSIGH.
1087
1088Protect against THB=0 in PYRAND.
1089
1090Protect against YSTMAX-YSTMIN = 0 in PYSIGH.
1091
1092Check for moved leptoquark at beginning of PYRESD just like for
1093other particles with colour.
1094
1095-----------------------------------------------------------------