A complete description of the PYTHIA program would have to cover four aspects:
The PYTHIA program is continuously being developed. We are aware of many shortcomings, some of which hopefully will be addressed in the future. Mainly this is a matter of including new interesting physics scenarios and improving the existing ones, but also some cleanup and reorganization would be appropriate. No timetable is set up for such future changes, however. After all, this is not a professionally maintained software product, but part of a small physics research project. Very often, developments of the programs have come about as a direct response to the evolution of the physics stage, i.e. experimental results and studies for future accelerators. Hopefully, the program will keep on evolving in step with the new challenges opening up.
In the longer future, a radically new version of the program is required. Given the decisions by the big laboratories and collaborations to discontinue the use of Fortran and instead adopt C++, it is natural to attempt to move also event generators in that direction. User-friendly interfaces will have to hide the considerable underlying complexity from the non-expert. The PYTHIA 7 project got going in the beginning of 1998, and is an effort to reformulate the event generation process in object oriented language. Even if much of the physics will be carried over unchanged, none of the existing code will survive. The structure of the event record and the whole administrative apparatus is completely different from the current one, in order to allow a much more general and flexible formulation of the event generation process. A strategy document [Lön99] was followed by a first `proof of concept version' in June 2000 [Ber01], containing the generic event generation machinery, some processes, and the string fragmentation routines. In the next few years, the hope is to produce useful versions, even if still limited in scope. Due to the considerable complexity of the undertaking, it will still be several years before the C++ version of PYTHIA will contain more and better physics than the Fortran one. The two versions therefore will coexist for several years, with the Fortran one used for physics `production' and the C++ one for exploration of the object-oriented approach that will be standard at the LHC.