The document discusses challenges in software sustainability and productivity faced by the National Science Foundation (NSF). It notes that NSF typically only funds software projects for 5 years, though many projects require support for 20+ years. It also discusses issues like a lack of career paths for software-focused researchers, inconsistent incentives and credit systems, training needs, challenges of interdisciplinary work, and ensuring software portability and dissemination. While the NSF has made some improvements through programs like SI2, the document concludes that more work remains to be done to address these challenges and push academic culture to better support long-term software projects.

1. Looking at
Software Sustainability and
Productivity Challenges
from NSF
Daniel S. Katz & Rajiv Ramnath, NSF
CSESSP Challenges Workshop
October 14 2015, Rockville, MD

2. Context: NSF SI2 Program
• NSF provides funds to support PIs
to develop and maintain the
software that is needed for NSF-
funded research and education
projects
• Rajiv Ramnath will talk about SI2 in
the Federal Agency panel later
today

3. Challenges
• Funding models. NSF supports projects for up to 5 years; lifetime
of software projects can span 20+ years
• International funding. Software collaborations span countries, but
most funding agencies don’t
• Career paths for software-focused researchers. University
structure &academic culture rewards publications, not software
• Incentives, including credit. How should software be cited? How
are all software contributions recognized?
• Training. What software engineering practices work in science?
• Interdisciplinary work. Much software requires knowledge from
more than one field – this doesn’t fit our siloed system
• Portability. How to deal with changing hardware, middleware, and
languages?
• Dissemination. How do we document available software with
usage examples, strengths, weaknesses, and user experiences?

4. Lessons Learned
• Both funding agency and academic culture do
not change as quickly as scientific practices,
and that continued effort is needed to raise
these issues
• We have dealt with some of them internally at
NSF, through SI2 as a whole, and specifically,
through Institutes, EAGER awards, workshops &
hackathons, and are working towards additional
community workshops and national activities
intended to create awareness of the problems
and of successes and failures in attempting to
solve them

5. Conclusions
• Overall, we believe that the SI2 program has
been very successful to date, but that there is a
lot of work remaining, including carrying out the
institute process, and both leading and pushing
academic community to develop solutions to the
existing sustainability and productivity
challenges.
• Material taken from:
– Katz, Daniel S. and Ramnath, Rajiv.
Looking at Software Sustainability and Productivity
Challenges from NSF. arXiv.org, 17 August 2015,
http://arxiv.org/abs/1508.03348