The open source development of scientific algorithms was awarded a recognition prize by Nature Methods

date: 07.02.2020

category: Sporočila za javnost

 

Nature Methods is a journal that publishes new or significantly improved methods of research techniques in the natural sciences. With an impact factor of 28,467 it is one of the most cited journals in the world. A few days ago a paper titled SciPy 1.0: Fundamental Algorithms for Scientific Computing in Python was published. The paper gives an overview of the performance and development of SciPy 1.0 from 2001 to today, and highlights some of the latest technical features. Since its first publication in 2001, SciPy has become the de facto standard for the development of scientific algorithms in Python.

SciPy has over 600 active developers, is the base on which more than 6,000 other packages are based, more than 100,000 different repositories, and has almost 20 million downloads per year. SciPy is a numerical library, which is used in practically all scientific fields, such as mechanical engineering, physics, computer science, biology, chemistry, medicine, astrophysics, linguistics, economics, and geography. A large number of scientific publications and industrial or commercial applications are based on open source. The entire SciPy ecosystem, including Numpy, has made significant progress in large-scale data exploration and machine learning over the past decade (47% of all github.com-based machine learning projects are SciPy-based). SciPy has also been used in some recent research, such as gravitational wave analysis and M87 black hole imaging.

The publication in this prestigious journal is not a classic scientific contribution, but a tribute to the entire community of developers (co-authors of the article) and their enormous collective effort, which has a huge social and scientific impact. It is largely based on the free and leisure efforts of individual researchers and is not measured today in any metric for measuring scientific work. In the past, open source has hardly been funded at all (e.g. until recently SciPy received practically no financial support and was based solely on the voluntary work of developers); recently there has been a growing interest in donors and research institutions.

SciPy is proof that open source can significantly accelerate scientific development, and the publication in Nature Methods was a recognition of this contribution. The paper has 111 authors who were selected by a special invitation from more than 600 active developers. We are extremely proud that Prof. Janko Slavič, PhD from the Faculty of the Mechanical Engineering University of Ljubljana is the Slovenian representative to be named as one of the authors of the paper. The invitation and co-authorship is a recognition of his efforts, his knowledge and his time, which has channeled him into the so-called “open science” for more than a decade. Prof. Janko Slavič, PhD is also the author of the university textbook »Programiranje in numerične metode v ekosistemu Pythona« ("Programming and Numerical Methods in the Python Ecosystem”).

The field of mechanical engineering has developed rapidly in recent years and is becoming increasingly important, the development of custom numerical models, cloud applications, artificial intelligence, high-performance computing. Open source opens up unlimited development possibilities here. The Faculty of Mechanical Engineering UL is a leader in the development of open source in the field of structural dynamics. The Faculty therefore maintains a number of different repositories, all based on the ecosystem of the SciPy library. This year, the faculty is organising for the second time an international, top-class summer school on structural dynamics and open source development.

slavič

back to list