Discussions

News: JADE Project Reborn As Javolution And jScience

  1. JADE Project Reborn As Javolution And jScience (6 messages)

    The Java Addition to Default Environment (JADE) project has been reborn under a new name umbrella, namely "two new projects with additional features and capabilities: Javolution (javolution.org) and jScience (jscience.org)".
    Because of trademark issues we had to rename our Java Addition to Default Environment (JADE) project. We did a little more than that, we created two new projects with additional features and capabilities: Javolution (javolution.org) and jScience (jscience.org).

    Java developers, please update your bookmarks. You may also read the 'Top 10 reasons' why you should consider using Javolution in your current Java project or how you can take part in this immense undertaking that the jScience project represents

    Javolution - http://www.javolution.org
    Transparent object recycling - No memory allocation, memory defragmentation or garbage collection ever.
    Configurable/Inheritable XML serialization/deserialization for any class.
    Allow your application/library to run on any platform from the J2ME CLDC 1.1 to the latest J2EE 1.5
    Text manipulation several order of magnitude faster than using java.lang.String.
    Guaranteed response time. No internal table resize/copy ever performed and for thread running in PoolContext no garbage collection either.
    Low-level concurrency support.
    The fastest implementations of the most common interfaces (e.g. FastMap for java.util.Map interface).
    Free...

    JScience - http://www.jscience.org
    A mathematics package, which includes a first (and I believe unique) generic Matrix class capable of resolving linear system of equations involving any kind of elements (e.g. Real, Complex, Quantity, Function, Vector, etc.).
    Functions package for symbolic calculations and analysis.
    All different types of numbers such as real numbers of arbitrary and guaranteed precision, or the always exact rational numbers.
    Units package (e.g. SI units) developped in collaboration with the JSR-108 experts group.
    Strongly typed quantities classes (physics package) with dimension-checking done in the form of class-type checking (e.g. the division of a Length by a Duration returns a Velocity instance). Quantities also take into account measurement errors as well as numeric errors (interval arithmetic).
    Support for Standard, Relativistic, High-Energy, Quantum and Natural physical models.

    Regards,

    Artem D. Yegorov
    http://www.activexml.org
  2. Why is this ignored ?[ Go to top ]

    I've seen mr. Dautelle for a year now actively pushing JADE.
    Forums, mailing lists, sun's bug parade and so on.
    I for one think some of JADE's claims, if not the work itself - which I've yet to try out - are simply amazing and for the moment completely elude me.
    Can it be that the things it claims it does are simply beyond belief that there seems to be a void of sound around JADE ?

    Are you using it ? Have you checked it out at least ?

    Regards
  3. Why is this ignored ?[ Go to top ]

    I've looked at some of the Javalution stuff and I'm definitely liking the jScience package. Some claims are indeed overbloated, but there are decent things about the framework and it definitely has its niche in extending capabilities of core Java.

    I would have to spend considerable more time in analysis and research to give out a more pronounced and valid opinion though...

    Just my 2 cents...

    Regards,

    Artem D. Yegorov
    http://www.activexml.org
  4. E-science really gets me hot! And I like jScience's mission, "To create synergy between all sciences (e.g. math, physics, sociology, biology, astronomy, economy, etc.)" That said, jScience seems more a wishful placeholder than a usable library. There's no fuel without solvers, Fourrier transforms, Gaussians, particle-in-cell, fluid dynamics, stochastic optimizers, etc.

    But more fundamental is the failure of Java to gain traction in numerics generally. JavaGrande.org seems to be going nowhere. Java's performance lags at numerics and multidimensions by an order of magnitude. Of course Moore's Law immensely favors Java, but e-scientists refuse to abandon their highly optimized ForTran and C++ legacies.

    Occasionally I could pick up Java GUI work for the federal labs, but Java almost never gets a chance at the underlying formulas. I think there's unnecessary bigotry against virtual machinery in the quantitative sciences, so I quit. I think the bigotry is an anti-competitive consequence of science being unnecessarily funded by Uncle Scam and CERN. If there were a free market for hard science, it would be a much better scene, and Java would be a key player on the quant boxes.
  5. This is only a start. Few developpers are currently working on sciences plugins/packages such as physics.nuclear, mathematics.integrations, computing.geneticAlgorithms, cognitive.neural-networks...
    Obviously, we need more and you are welcome to participate.
    The amazing thing I can tell you about the upcoming plugins is how well they integrate and take advantage of each others in a variety of ways. The synergy has already started and it is fascinating...
  6. Few developpers are currently working on sciences plugins/packages such as physics.nuclear... The amazing thing I can tell you about the upcoming plugins is how well they integrate and take advantage of each others in a variety of ways. The synergy has already started and it is fascinating...

    Google reports that the word 'plugin' doesn't appear anywhere on jScience.org, so I understand now how premature TheServerSide is to announce it. Once the plugins are available, I'll know who to send my resume to. I really want to build e-science backends with Java, and until now I thought I was freakishly alone for choosing Java. The USA's Dept of Energy is pumping millions of dollars yearly into Java e-science, but it's all middleware and GUIs, neither of which challenges me. Really I want to do proteomic backends in Java, but alas little or no such development occurs in Colorado.
  7. Just a small piece of positive feedback for M. Dautelle's work. We've been using parts of the former J.A.D.E. for a little over a year now, and my view is that the claims made for the product are well-founded (at least as far as the parts we've used are concerned).

    For the record, we've made extensive use of the quantity/unit structure now in jscience, happily added additional measurement units, following the very well documented procedure to make sure the framework's structure is followed, and developed a time-series analysis framework from the jscience Matrix class. We've also successfully used some of the classes now in the javolution library, notably the 'real-time' collection implementations.

    In other projects, we have used other unit handling libraries, and we are planning to use the jscience library wherever using java is appropriate. So, thanks to M. Dautelle for the excellent work!