\title{MPP: A Framework for Distributed Polynomial
Computations\thanks{Work reported herein has been supported in
 part by the National Science Foundation under Grant CCR-9503650, by the
 Deutsche Forschungsgemeinschaft, and by the Stiftung Innovation des
 Landes Rheinland-Pfalz.}
}

\author{
Olaf Bachmann\thanks{Also: Department of Mathematics and Computer Science,
Kent State University Kent, Ohio, email: obachman@mcs.kent.edu.}
 ~~~~~~Hans Sch\"{o}nemann\vspace{2.5 mm}\protect\\
Centre for Computer Algebra\\
Department of Mathematics\\ 
University of Kaiserslautern\\
Kaiserslautern, Germany \\
\small{\sl \{obachman,\,hannes\}@mathematik.uni-kl.de}
\and
Simon Gray\vspace{2.5 mm}\protect\\
Department of Mathematics and Computer Science\\
Kent State University\\
Kent, OH 44242, USA\\
\small{\sl sgray@mcs.kent.edu}
}
 
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\begin{abstract}
There are obvious advantages to providing communication links between
independent software tools, including the ability to do parallel
distributed computation, distributed problem solving, and providing
more direct access to a wealth of computational resources.  The
challenge of providing {\sl connectivity} is to produce homogeneity in
a heterogeneous environment.  We have explored this problem within the
context of applications specially designed for polynomial
computations.  Our solution uses the Multi Protocol (MP) to establish
communication links between independent packages, and the ideas of
dictionaries, prototypes, and annotations provided in MP.  We describe
the design of an MP dictionary for polynomial computations ({\sc
  mpp}), as well as the implementation of interfaces to {\sc
  Singular}, {\sc factory}, and Mathematica based on this dictionary.
Important aspects of the design and implementation include generality,
efficiency, and the ability to convey supplemental information.
  We include a discussion of our experiences and some timings.
\end{abstract}

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%