User:Gnu andrew
From Wikipedia, the free encyclopedia
[edit] Academic Life
I'm currently a first year postgraduate student at the University of Sheffield, looking at the combination of timing and mobility with respect to process calculi. The field of process calculi is concerned with the modelling of concurrent systems, via their representation as interacting processes. My particular interests lie in algebraic process calculi, where the processes and their interactions are modelled using algebraic notation with appropriate semantics. The calculi I work with have a common heritage in Milner's CCS, which utilises the basic operations of parallel composition, non-deterministic choice and recursion on processes, which interact via synchronizing on channels.
While CCS is suitable for the local synchronization of two processes, it has problems scaling to systems of an arbitrary size in a compositional manner. My calculus of Typed Nomadic Time (TNT) is based on CCS, and attempts to combine abstract time, distribution and mobility within a single algebraic system. It allows the creation of a compositional semantics for mobile component-based systems, which utilise the notion of communication between arbitrary numbers of processes within a dynamic topology.
With abstract time, as seen in TPL, PMC, CSA and CaSE, we can perform global synchronization between any number of processes. For example, this allows us to broadcast a message without making the sender dependent on the number of receivers. In pure CCS, moving from one recipient to two, for example, requires either changing the sending process so that it broadcasts the message twice, or creating a recursive process, which can't be deterministically stopped. Adding clocks that respect maximal progress (allowing all synchronizations to take place before a clock tick) allows us to build a recursive process that stops when the clock ticks (i.e. no more synchronizations can take place).
To this, I add a notion of distribution and mobility, as seen in Cardelli and Gordon's ambient calculus and its variants. This gives the calculus both a notion of distribution or spaciality, as well as a more natural form of mobility, compared to the scope mobility of the π calculus. Additionally, localities seem to provide a certain synergy with the notion of hierachically scoped time.
[edit] Free Software
I'm an active member of the Free Software community. I'm an associate member of the FSF and a developer on the GNU Classpath project.