Just how far can Java be pushed in the micro-embedded space? That's what Eric Arseneau, principal investigator of the Squawk virtual machine and his team want to know. With the goal of being “as invisible as possible”, in order to enable multitudes of companies to do multitudes of things, Arseneau is looking beyond Sun SPOTs to targets devices like SIM (Subscriber Identification Module) cards and similar devices running approximately 16K of Flash and 2K of RAM.

Vice President of Global Systems Engineering, Hal Stern, welcomes Arseneau to this latest edition of Innovating@Sun to delve deeper into the possibilities in the embedded world thanks to technology such as Project Squawk. The interview includes:

  • Goals for this small-scale “Java operating environment” - as Eric Arseneau likes to call it
  • Why Sun open sources both the Squawk virtual machine and the hardware specifications for what it can run on
  • Efforts to not only expand the marketplace, but create licensing opportunities as well
  • Moving beyond traditional computer science to target cross-disciplinary applications among audiences such as universities, artists, and scientists to name a few
  • The expected proliferation of very small, embedded devices given a lower barrier to entry for developers and manufacturers
  • Why future embedded devices will have increased orders of magnitude over the cell phone

    Links:



    Transcript
    Squawk Operating Environment



  • Fortress 1.0, Sun's new programming language for high-performance computing, is all about opening up scientific domains for writing code – domains that have previously been inaccessible to the average programmer. Dr. Eric Allen, principal investigator in the Programming Language Research Group at Sun Labs, joins Hal Stern on this latest edition of Innovating@Sun to elucidate the features and benefits of Fortress. Allen and Stern discuss:

  • Fortress' design focused on high programmer productivity
  • Innovative features including syntax and parallelism and why these will be useful for targeting multi-core architectures as well as super computers
  • How the mathematical notation in Fortress decreases the space between what the specification of a program looks like and what the corresponding code looks like, thereby reducing errors
  • Dynamic compilation and designating optimizations you want the compiler to make
  • Why much of the language is put into the libraries
  • Dimension checking and the typing system
  • Why transactions are built into the core of Fortress
  • Balancing the need for correctness vs. allowing the developer to go back and make fixes vs. being fast and scalable
  • Open source, communities and getting Fortress into the hands of developers

    Listen to the entire podcast to hear why Sun sees Fortress as its best attempt to design a programming model that meets the needs of the 21st century programmer in terms of multi-core architectures and parallelism requirements.

    Links:



    Transcript
    Project Fortress