ESIP Software Sustainability Session
Bob Downs started off by answering some questions about activity undertaken at Summer Meeting 2014:
"Why seek ESIP perspectives on Science Software Sustainability?" While the ESIP community has been successful at articulating best practices around data management, software has been relatively unaddressed. This is surprising, given the fact that software and issues in data management, such as data quality, reproducibility, etc. are tightly coupled. So, a first step in developing best practices in software stewardship should include a base-line assessment of what the ESIP community's best practices are, and what they see as important. Essentially, we asked "how do we treat and sustain software that makes that make data meaningful?"
Next, Chris offered findings from the exercise, as described in the following paper (http://commons.esipfed.org/node/2633) and set of slides (http://commons.esipfed.org/sites/default/files/LenhardtSoftwareClusterSession.pptx).
Joe Hurcle - On first recommendation - The best way to get science input is to go to them....
We should agree upon and promote a group definition of sustainability. Further, we should - in a basic way - say what we mean when we say software. What are the boundaries of this group? Finally, we should use these definitions to identify ways that we can support that with software.
- One suggestion is to look at work from XCEDE workshops, and model best practices as articulated by HPC community.
Decadal survey should include some discussion of software engineering priorities
Next Telecon session
- January 16th - 2pm EST.
- Software Carpentry - we should decide whether or not this is a session we would like to support. - Use cases we might like to see this group take on for summer session (around documentation, variable names, etc) - Organize ourselves to move towards a short course in software.
For next meeting
Nic will develop a template for use cases (i.e. what are the metadata that we want to collect)
- The problem
- The ESIP setting for the problem
- What is a sustainable software practice - what is ideal and what works and what does not?