Summary

Respondents experiment at an average of two facilities each while using an average of 3.4 operating systems, by necessity. Users are generally frustrated by VMS-based data formats at ISIS, PSI and KEK which must be converted to port to other operating systems and which FTP poorly to VMS machines. They try to use their favorite analysis programs with data from different facilities, if they can convert the data. Sometimes, more than one program is used because not all theory functions are available in a particular program. Often, facility programs can only be used on the facilities' VMS computers because versions are not available that will run under Windows or Linux. Porting of software from VMS is particularly problematic for software from TRIUMF. Programs from KEK are seen as most in need of updating.

Users summarize and ``meta-analyse'' their results using an average of 2.3 math programs each, with a wide variety of programs used. Those improving facility software should take this into account, when designing the output of results.

Users are equally supportive of the ideals of a universal data format and of a universal suite of facility programs.