Banach, R., Balazs, J. and Papadopoulos, G.A. (1995) A translation of the pi-calculus into MONSTR.  Home/Search   Document Details and Download   Summary   Related Articles   Check

.... state] 2] Counter[self state] self: Channel Empty , Assign[ans chan state] Counter Match[self:Channel[a] Inc[value ans chan] state] 3] #Counter[self newstate: ADD[value state] 12 self: Channel Empty, #Assign[ans chan newstate] Counter Match[self mess state] [6] = ##Counter[ self mess state] That this works as required relies critically on the fact that each channel node has a unique owner, this being the Counter or Counter Match function node in this case. Because the only other nodes allowed to rewrite a channel node are Send functions, and ....

....activities of the generated graph rewrite rule sets. This is a traditional and extensively tested approach, having been used for a variety of language for 16 malisms ( 7,8] A full translation gives the added benefit of allowing rigorous reasoning about all aspects of a language; e.g. the way [6] gives an alternative perspective on process calculi. And there is no reason why we could not apply the principles of our approach to other similar families of languages such as the concurrent constraint ones ( 10] ....

R. Banach, J. Balazs and G. A. Papadopoulos, "A Translation of the Pi-Calculus into MONSTR", Journal of Universal Computer Science, Springer Verlag, Vol. 1, No. 6, 1995, pp. 335-394.

.... models among themselves as well as with MONSTR ( 16] Another is that one can reason about the behaviour of some model or language by transposing the discussion to the level of MONSTR where the rewriting and interaction (in the general sense) between processes or agents becomes more explicit ([8,11]) Finally, MONSTR provides a natural implementation apparatus for a variety of such computational models and languages ( 2,9,10,11,26] In the context of the present work, the mapping of Interaction Nets onto MONSTR serves, among others, two purposes: to provide an implementation apparatus for ....

....of agents to synchronise at a port node in order to interact, then one faces a much harder implementation problem, as agents can no longer be allowed to grab a port node eagerly. Similar issues arise when one wishes to take seriously the synchronisation model inherent in process algebras (see [8]) Thus, while in relaxing the port invariant 8. In the latter case presumably the port node in the middle of such a LHS edge cannot be regarded as garbage after the rewrite. 9. For binary port connections and agents having single principal ports, complementary types force injectivity of the port ....

Banach R., Balazs J., Papadopoulos G. A., "A Translation of the Pi-Calculus into MONSTR", Journal of Universal Computer Science 1 (6), 1995, pp. 339-398, http://www.iicm.tu-graz.ac.at/jucs.

....this paper is effectively a parallel (distributed) realization of the Interaction Net formalism. The rest of the paper is organized as follows. The following section describes MONSTR and some of its more important properties. Section 2 does the same for Interaction Nets, following the treatment of Banach (1995). The two models are brought together in section 3 which describes a translation of typed Interaction Nets into MONSTR. Section 4 shows how untyped Interaction Nets may also be translated. Section 5 presents the translation to equivalent MONSTR rule systems of some concrete Interaction Nets ....

....Nets were invented for describing fine grained computations graphically. Their theory builds on prior work in multiplicative Linear Logic that gives the Interaction Net model particularly transparent properties regarding confluence, and to a lesser extent normalization. We use the formulation of Banach (1995) as it is more convenient for the translation that we subsequently give. Interaction Nets can be viewed as bipartite graphs where the two node kinds are agent nodes and port nodes. Each agent bears a symbol, which determines the number of port edges incident on it, and the attributes of those ....

[Article contains additional citation context not shown here]

Banach, R., Balazs, J. and Papadopoulos, G.A. (1995) A translation of the pi-calculus into MONSTR.

....able to accommodate the needs of, often divergent in behaviour, computational models; needs that range from those associated primarily with reasoning and specification to those more related to implementation issues. In [5] it is shown how concurrent logic languages can be implemented in MONSTR. In [4] MONSTR is used to reason about pi calculus and in [7] it is shown how MONSTR can be used as an implementation and specification framework for concurrent object oriented languagues. Finally, 6] studies the possibility of using MONSTR to implement concurrent languages based on Linear Logic but ....

Banach R., Balazs J., Papadopoulos G. A., "A Translation of the Pi-Calculus into MONSTR", Journal of Universal Computer Science, 1 (6), 1995, pp. 339398, http://www.iicm.tu-graz.ac.at/jucs.

....the MONSTR computational model. We then discussed a number of issues pertaining to synchronisation of concurrent method invocations between and within an object by lifting the relevant discussion from the level of a COOL to that of a set of MONSTR rewrite rules. This generalises our initial work (Banach and Papadopoulos 1995b) where we focused on the object oriented functional language UFO (Sargeant 1993) It is important to note that our simple object abstraction renders the current work independent of any particular COOL. This means that MONSTR provides good neutral ground for comparing different synchronisation ....

....framework for COOLs by developing fully fledged language translators to MONSTR, where emphasis is on optimising the run time activities of the generated graph rewrite rule sets. This is a traditional and extensively tested approach, having been used for a variety of language formalisms (Banach and Papadopoulos 1993, 1995a) A full translation gives the added benefit of allowing rigorous reasoning about all aspects of a language; e.g. the way Banach et al. 1995) gives an alternative perspective on process calculi. And there is no reason why we could not apply the principles of our approach to other similar ....

[Article contains additional citation context not shown here]

Banach R., Balazs J., Papadopoulos G. A. (1995), A Translation of the Pi-Calculus into MONSTR, Journal of Universal Computer Science, 1 (6), 335-394.

....help with this, by allowing executions that differ only by sequences of permitted interchanges to be regarded as equivalent. Ultimately one may be able to reduce all the executions to a standard form if the understanding of the system is profound enough. For a concrete example of this process see [Banach et al. 1995)] Actually, to just say that independent actions may simply be interchanged is too simplistic. In one case, choosing one action first obliterates the other; in another, one choice entails an extra action. And in any case, the results we start with in this section are expressed in a symmetric form ....

Banach R., Balazs J., Papadopoulos G., A Translation of the Pi-Calculus into MONSTR. J.UCS 1, 335-394, (1995).

....language, MONSTR subsequently proved to be a flexible and expressive model of computation in its own right. For a variety of applications in addition to its use in the Flagship Machine, see e.g. Banach and Papadopoulos (1993) Banach and Papadopoulos (1995a) Banach and Papadopoulos (1995b) [Banach et al. 1995)] Banach and Papadopoulos (1996a) Banach and Papadopoulos (1996b) Given the retrenchment to the MONSTR sublanguage, the original question then became that of the correctness of the MONSTR implementation. The question is non trivial since the semantics of the language and of the machine ....

Banach R., Balazs J., Papadopoulos G., A Translation of the Pi-Calculus into MONSTR, J.UCS 1, 335-394, (1995).

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC