The implementation of BousiProlog by means of a meta-interpreter
is a cheap solution but expensive from the point of view of the
efficient execution. In order to solve this efficiency problem,
we have investigated how to incorporate the weak unification
algorithm into the Warren Abstract Machine (WAM).
It is convenient to say that, at the present time, the Similarity
WAM (SWAM) does not implement all the features of full BousiProlog. It is a
prototype implementation useful to essay new compilation techniques.
In the future we want to develop the SWAM to cover
all the present and future features of Bousi
Prolog in a more efficient implementation.