理论与计算化学程序软件发展需求与资助模式思考——基于美国自然科学基金委软件基础构架项目

On the need and funding model for supporting theoretical and computational chemistry software development:aperspective based on the software infrastructure for sustained innovation program of the US National Science Foundation

随着计算机计算能力和算法的高速发展,现有的计算软件已不能满足目前日益增长的需求。因此开发一种实用、高效的新型的计算软件结构就成为当务之急。该新型结构将具有支持多层理论方法的能力,同时可以兼顾实验和理论化学的结果,优化搜索引擎。新一代的计算化学软件必须具备高度的并行性、可扩展性、重复使用性和兼容性,并可实现大型社区驱动。本文基于美国自然科学基金委软件基础构架项目,阐述了计算化学软件发展需求和相关的资助模式,包括层列式的资助模式和资助范围。此外,本文还将从开发软件项目的特点等方面展开讨论,包括软件基础构架,软件培训和推广,以及其对软件开发项目可持续性的影响。

With recent advances in computational power and algorithmic efficiency, we cannot rely on incre- mental changes to existing computational chemistry software. A new research infrastructure is needed that smoothly incorporates multi-scale methods into a cohesive experimental and theoretical chemistry discover- y, optimization and design engine. The next generation of computational chemistry software for treating re- alistic chemical systems with atomic resolution will need to be highly parallelized, extensible, reusable, in- teroperable and community-driven. To this end, this article aims to present a perspective on the need and related funding model for supporting computational software development based on the Software Infrastruc- ture for Sustained Innovation Program of the US National Science Foundation. A tiered model and its re- spective funding scope are illustrated herein. Important considerations that focus on software infrastruc- ture, education and dissemination, and their impact on the sustainability of a software development pro- gram are discussed in this article.