Talk π`s in Debian or Scientific Debian: NumPy, SciPy and beyond

Presented by Yaroslav Halchenko (Debian, NeuroDebian, PyMVPA) in Scientific track 2011 on 2011/08/27 from 16:30 to 17:00
Abstract

Debian project delivers a complete free and open-source operating system suitable for deployment virtually everywhere: from mobile phones to high-performance computing centers. Because of the well known stability and versatility, Debian has not only became the origin for more than a hundred derivative Linux distributions, such as Ubuntu: Together with the extensive coverage of scientific packages, Debian is considered by many to be the best available scientific platform.

Due to the beauty, versatility, and wide adoption of Python, it is not surprising that multiple Debian teams and developers focus on packaging and maintaining of Python-based software. As the result, with over 140 packages depending on NumPy, over 30 on SciPy, and multiple officially simultaneously supported versions of Python, Debian becomes "the Python distribution" providing the largest capabilities for the numerical computing in Python out of the box. Such a turnkey solution eliminates majority of the problems often encountered by users trying to deploy Python-based analysis workflows on their systems. With clean description of build- and run-time dependencies, package build-time testing, and generic QA (e.g. frequent rebuilds, transition testing) provided by Debian project, scientific software in Debian obtains additional assurance toward robust performance after its installation on users' systems.

This talk will cover

  • benefits and possible limitations of doing scientific computing using Python "the Debian way"
  • the current status of Python in Debian: 2.x and 3.x series, upcoming transitions, wheezy release goals
  • packaging and maintenance of Python modules in Debian: common practices, existing standards and procedures
  • existing scientific Python applications in Debian: coverage of scientific domains, aspects of integration with other scientific packages
  • reasons and channels to contribute to Debian

This talk should be of interest both for the scientific Python developers to gain insights on possible benefits from the integration of their software in Debian, and for the Python "users" to realize what comes within the Debian box.

tagged by
no related entity