breakout room 3

[alee]

Dawn Parker, Gunnar Dressler, @platipodium @benbovy @chrstngyn

The existing preamble is tailored towards reproducibility and replicability. This should be revised to fit the theme which is reusability of model code.

• come up with a way to relate the three R's
• revise the goal statement to reflect the focus on reusability
• replace technical jargon with clear, simple language (references to containerization, continuous integration etc)
• consider leaving verify in the Goal
• OSI approved license is not requirement for reuse, instead just a license that allows for reuse
• relationship between maintenance of the code and reuse

Issues with monolithic code and its impact on reusability. We need to clarify the scope as this document is heavily tailored towards model coupling and component based coupling.

[platipodium]

See disambiguation between replication and reproducibility: https://en.wikipedia.org/wiki/Reproducibility

[benbovy]

Some thoughts on reproducibility vs. reusability in this blog post: http://gael-varoquaux.info/programming/beyond-computational-reproducibility-let-us-aim-for-reusability.html

[platipodium]

In Earth System Modeling, there's alwa the issue on bitwise reproducibility from simulation models

[platipodium]

Some thoughts on reproducibility vs. reusability in this blog post: http://gael-varoquaux.info/programming/beyond-computational-reproducibility-let-us-aim-for-reusability.html

See their headeing "Beyond computational reproducibility, let us aim for reusability"

[benbovy]

One important aspect I see in "reusability" is maintaining software over the years. For example, we might be able to use some old software to reproduce the results from a related publication, but we might not be able to leverage this software anymore for doing new research because the whole software ecosystem on top of which it as been built as evolved too much / too rapidly.