The Apache Foundation did the math. The end result is version 2.0 of the Apache Commons Math library, which was released this past weekend. The new version of this open-source math library brings in better support for linear algebra, new optimizations and genetic algorithms.
Phil Steitz, committer on Apache Commons Math and an Apache Software Foundation member, was the release manager for 2.0. He said that a big focus for this version was improving the API for the library. That meant breaking backward compatibility, however.
“[For 2.0], we decided that we really wanted to increase the required JDK level to JDK 1.5," he said. "The Apache Commons Math project started in 2004, and a lot of the original code was written to run on even 1.2. We decided in 2.0 that we wanted to fully modernize the codebase by doing things like bringing in support for generics.
"The library had grown significantly, and to make it more accessible and address some problems we had with the early versions of the API, we decided we wanted to repackage some of the code to address some of the mistakes we made in the early version of the API."
Thus, version 2.0 of the Apache Commons Math library requires JDK 1.5 for the first time.
Another major focus for this release was linear algebra. In 2004, when Steitz contributed the first bits of code that formed the Apache Commons Math project, there weren`t many options in the open-source math library world of Java, said Steitz.
As an example, he called out JAMA, the Java Matrix Package, originally created by the National Institute of Science and Technology. JAMA, while useful, was not being maintained in 2004, he said, and that meant there were no optimizations being done on the code.
The JAMA API was an influence on the Apache Commons Math library API, said Steitz. He added that, with the inclusion of the new support for linear algebra, sparse matrices and matrix decomposition algorithms, the Apache Commons Math library has filled in the gap left open by the long-ignored JAMA.
Steitz said that, over the past year, the Apache Commons Math library has enlisted more than 35 contributors. Those contributions are helping to push the project ahead, and he said he expected the new genetic algorithms to be a big focus in the next year.
The Apache Commons Math library is available at the Apache website.
Source:
http://www.sdtimes.com/APACHE_COMMONS_MATH_LIBRARY_HITS_2_0/By_Alex_Handy/About_
Phil Steitz, committer on Apache Commons Math and an Apache Software Foundation member, was the release manager for 2.0. He said that a big focus for this version was improving the API for the library. That meant breaking backward compatibility, however.
“[For 2.0], we decided that we really wanted to increase the required JDK level to JDK 1.5," he said. "The Apache Commons Math project started in 2004, and a lot of the original code was written to run on even 1.2. We decided in 2.0 that we wanted to fully modernize the codebase by doing things like bringing in support for generics.
"The library had grown significantly, and to make it more accessible and address some problems we had with the early versions of the API, we decided we wanted to repackage some of the code to address some of the mistakes we made in the early version of the API."
Thus, version 2.0 of the Apache Commons Math library requires JDK 1.5 for the first time.
Another major focus for this release was linear algebra. In 2004, when Steitz contributed the first bits of code that formed the Apache Commons Math project, there weren`t many options in the open-source math library world of Java, said Steitz.
As an example, he called out JAMA, the Java Matrix Package, originally created by the National Institute of Science and Technology. JAMA, while useful, was not being maintained in 2004, he said, and that meant there were no optimizations being done on the code.
The JAMA API was an influence on the Apache Commons Math library API, said Steitz. He added that, with the inclusion of the new support for linear algebra, sparse matrices and matrix decomposition algorithms, the Apache Commons Math library has filled in the gap left open by the long-ignored JAMA.
Steitz said that, over the past year, the Apache Commons Math library has enlisted more than 35 contributors. Those contributions are helping to push the project ahead, and he said he expected the new genetic algorithms to be a big focus in the next year.
The Apache Commons Math library is available at the Apache website.
Source:
http://www.sdtimes.com/APACHE_COMMONS_MATH_LIBRARY_HITS_2_0/By_Alex_Handy/About_
