Hopefully this piece of news will be of interest to you: prompted by the unprecedented accuracy of Kepler light curves, the PHOEBE development team is working on a complete redesign of PHOEBE. We are refining all aspects of the code to come up with a flexible, modern framework that will (1) meet the requirements for ultra-high precision modeling, and (2) allow for easy and convenient expansion of the feature-set. Our development motto is that PHOEBE should be able to model a UFO eclipsing a Ferrari.
As an example, you may click on the image above (or this link) to see a small animation of triangulating the surface of a star by the newly implemented marching method based on the work by Hartmann (1998), The Visual Computer, 14, 95. The algorithm performs robustly on distorted surfaces (i.e. overcontact binaries) as well. This will significantly improve the accuracy of the model. For the impatient, the new code resides in the sandbox area on SVN. Interested in details or joining in on the effort? Drop us a line on phoebe-devel.
A quick note that documentation is now available in pdf. You may view and download it from here. A current snapshot of the manual has also been uploaded. It still requires lots of work, but at least the work started. As always, all help is appreciated.
Artificial Intelligence Approach to the Determination of Physical Properties of Eclipsing Binaries: The EBAI Project
In the time of fully automated sky surveys that produce terabytes of data at incredible rates, modeling engines must evolve accordingly. Our attempt to address this challenge is to resort to artificial intelligence, namely back-propagating neural networks.
EBAI source code, along with the polyfit pre-processor, are released under GNU GPL and are available for immediate download from the Plugins and extensions page.
After two weeks of bug hunt, PHOEBE development team is proud to release PHOEBE 0.31a! We would like to express our utmost gratitude to everyone that took the time and reported the issues with PHOEBE 0.31 to us -- without them catching so many problems would not have been possible. At the same time we would like to encourage everyone to report any glitches and to post any questions on our mailing lists.
