Minutes of Event Generators Group, Orsay, 6th April 1998

Hywel started by telling us the parable of the two Monte Carlo authors. The moral: it is pointless to write a super-duper state-of-the-art program, unless it is flexible enough and well enough documented to be used by experimenters for event simulation.

On the whole, the status of dedicated generators is pretty good: we are already at the stage where most processes of interest have at least one generator. However, in many cases ingredients are missing, such as beam polarization and standardized interfaces.

It was pointed out that almost nothing exists for simulating R parity violating SUSY processes.

The group identified three areas where interfaces will be important:
beamstrahlung. Here a standard program by Thorsten Ohl exists. All available programs should be interfaced to it.
fragmentation. Programs should output events, preferably unweighted, in a standard format, so that programs like PYTHIA and HERWIG can add parton showers and hadronization in a standard way. We agreed that this should be based on the HEPEVT standard, but this is not specific enough. Issues of colour structure have to be addressed, and perhaps spin structure (eg for tau production). Some of these are the same as for four-fermion generators for LEP2, but not all, so we need to generalize the standards used there. No agreement was reached. This is an important topic for the next meeting, as serious event simulation cannot start until interfaces have been set up.
detector simulation. The hadron-level results of event generators should be output in a standard format for input into detector simulation. It was agreed that this should be via an ascii file. For the final-state hadrons themselves, this is relatively straightforward, but agreement still needs to be reached on how much internal information (="Monte Carlo truth") needs to be passed on, and how to do this. A draft standard should be written and circulated for discussion at the next meeting.

The perennial problem of providing a standard set of tuned parameters for PYTHIA and HERWIG was discussed. No agreement was reached. Hywel suggested that if no standard set can be agreed, the DELPHI set should be used. There was a little grumbling, but no concerted opposition.

For top events near threshold, accurate simulation of the final state is important. There exists a dedicated generator: TOPIK. It should be made publicly available during the workshop, and interfaced to PYTHIA and HERWIG. It could then also be used to study how accurate they are.

On the question of `standard processes to benefit from extra study', we essentially agreed that all physics processes should receive extra attention. More guidance from the physics working groups is needed here.

Another summary, together with the full range of `standard processes' discussed can found in Hywel's summary slides from the plenary session.