| Popis: |
Software development in the Computational Sciences has reached a critical level of complexity in the recent years. This “complexity bottleneck” occurs for both the programming languages and technologies that are used during development and for the infrastructure, which is needed to sustain the development of large-scale software projects and keep the code base manageable [1].As the development shifts from specialized and solution-tailored in-house code (often developed by a single or only few developers) towards more general software packages written by larger teams of programmers, it becomes inevitable to use professional software engineering tools also in the realm of scientific software development. In addition the move to collaboration-based large-scale projects (e.g. BrainScaleS) also means a larger user base, which depends and relies on the quality and correctness of the code.In this contribution, we present the tools and infrastructure that have been introduced over the years to support the development of NEST, a simulator for large networks of spiking neuronal networks [2]. In particular, we show our use of• version control systems• bug tracking software• web-based wiki and blog engines• frameworks for carrying out unit tests• systems for continuous integration.References:[1] Gregory Wilson (2006). Where's the Real Bottleneck in Scientific Computing? American Scientist, 94(1): 5-6, doi:10.1511/2006.1.5.[2] Marc-Oliver Gewaltig and Markus Diesmann (2007) NEST (Neural Simulation Tool), Scholarpedia, 2(4): 1430. |
