I would like to bounce an idea of the vcampus; something that I've been think about for awhile but haven't really done anything about.
The PE (performance equations) and totalisers are two really powerful PI systems that cover a large number of the base calculation requirements; IMHO you shouldn't use ACE and analysis rules for everything. The one disadvantage with PE's and Totalisers is the lack of integration with element templates; you can't define the formula within a template and create instances of the calculation.
Therefore, my thought was would it be practical to create a new data reference that contains the link to the point and the point configuration? The DR would allow the building of formulas in a similar manner to the formula DR but would have the added portion of a PI Point DR. When you create an instance of the DR the formula would be resolved to the source PI points and be written to the extended descriptor of the target PI point. There would naturally have to be a fair amount of checking code to ensure that the source attributes are indeed PI Point DR's or "static" values; no other DR's would be possible. This leads to the natural question of whether it is possible to inherit a DR from an existing DR? The way I would create this is to extend the PI Point DR to include a section for the formula, and reading/writing the configuration strings (both the AF and PI Point exdesc strings). I seem to recall an AF limitation that an assembly must be stand alone.
Is this practical? Worthwhile looking at?