Implicitly Distributing Pervasively Concurrent Programs
Distributed programs are often written as a collection of communicating modules. For example, to use Java RMI, programs are divided into objects which can be remotely referenced. Yet, in many cases, it would be desirable to write the program without the program structure being driven by distribution decisions. If distribution is decoupled from program structure, the programming language must allow communication throughout a program’s structure, instead of at a few known points. This situation is simplified if the programming language provides a uniform programming model for local and remote values (location transparency). We present Distributed Orc, which offers location transparency, and distributes program operations automatically in cooperation with the execution environment. By eliminating any special semantics of remote values, Distributed Orc enables programmers to write cohesive distributed programs, rather than programs artificially divided at distribution boundaries. Distributed Orc is derived from the Orc language, a (centralized) concurrent orchestration language.
Sun 17 JulDisplayed time zone: Amsterdam, Berlin, Bern, Rome, Stockholm, Vienna change
10:00 - 12:30 | |||
10:00 30m | Implicitly Distributing Pervasively Concurrent Programs PMLDC John A. Thywissen The University of Texas at Austin, Arthur Michener Peters The University of Texas at Austin, Texas, USA, William Cook UT Austin Link to publication DOI | ||
10:30 30m | Static Type Checking for the Kompics Component Model -- Kola - The Kompics Language PMLDC Link to publication DOI File Attached | ||
11:00 30m | Towards Aggregate Programming in Scala PMLDC Link to publication DOI | ||
11:30 30m | Building Distributed Applications with Data and Time Frames PMLDC | ||
12:00 30m | Fractional Permissions for Race-Free Mutable References in a Dataflow Intermediate Language PMLDC |