The current PaPCo philosophy is to keep all data for all panels in memory so that they can be plotted and manipulated together fast. The idea here is to encourage the user to use PaPCo for analysis too, where speed is required for zooming, slicing etc. - PaPCo for this reason provides to the user a variable called IsRead, which is set whenever the data for a given module is in memory, to let the part of the user-module that does the plotting know that the data does not have to be read again.
This does, however, assume that memory is not a problem. When plotting large
arrays of data for long time periods for many panels, the amount of memory
required might be very large. PaPCo makes no attempt at optimizing memory
usage: The basic idea is that memory is cheap. PaPCo has been developed on a
SUN SPARC 20 Station with 96 Mbyte of memory and no serious slow down due to
swapping has been encountered so far. In practice, it makes sense to close
down other applications for a serious PaPCo session, and to make sure that
the swap space is adequate (we use swap=memory 
4).
Eventually, however, it makes no sense in returning 10 days worth of 6sec resolution data, since no screen can plot this in a way which allows one to see all that resolution. It is again the responsibility of the user to write ``clever'' read-routines, which automatically reduce the resolution of the data for long time intervals.