The Simulated Package Principle (SPP)

    

The mapping by renaming of functions as explained in the previous sections can only work properly if each function carries a tag  that identifies, for which domain of arguments the function can be applied. This mechanism is similar to the package principle in LISP,  except that tagging does not happen automatically but must be done by the programmer. We call it therefore the Simulated Package Principle (SPP).

The name of the tag corresponds to the data type up to capitalization. (Syntactically, we append an abbreviation of the domain name as tag, separated by the underscore character ``_'').

By this mechanism we keep the domain independent of all its subdomains. Summarizing, the advantages of this independence are that

• we avoid unnecessary duplication of code,
• it is very easy to change from one implementation of a domain to another one, and
• the system can very easily be extended by new implementations for already existing domains, because the only thing necessary is to supply a coercion unit for the new implementation and make this active.

The last point is particularly important if the changes are supposed to be made on a very low level in a hierarchy  of domains (in our case, for instance, the domain of power products), because without SPP every change would have to be propagated until to the top of the hierarchy.