Workshop 4 Cyberinfrastructure Working Group discussion

[alee]

Breakout Room 3

Coordinate with certification group on tools for certification

Have others created standards and disseminated them?

Sponsors like Sloan are very much interested in maximizing the impact of their investments by having tools that improve the reach and longevity of their funded projects

Find ways to mesh with and curate funded integrated modeling projects (as opposed to bespoke wired together things)

reusability standards

attention on libraries of good examples for interoperability standards: sensible file formats, efficient checkpointing, temporal / control mechanisms, workflows.

compare / contrast workflow management packages for different use cases

provide cookbook/ case studies of different real workflow examples

DELTARES using BMI for instance

ruthlessly prioritizing things that save people time and have value

come up with exemplars

[cpritcha]

Templates and libraries to setup a project:

• https://usethis.r-lib.org
• https://drivendata.github.io/cookiecutter-data-science/

[cpritcha]

Repository automation around model certification, compliance and standards

Some bots as inspiration

• https://github.com/openjournals/whedon-api
• https://github.com/Mic92/nixpkgs-review (focused on package recipes hosted in a shared repository)

[alee]

Breakout Room 4

What are the responsibilities of the CYWG?

1. support certification: badges, evidence, automated tests
2. be the glue for systems of systems modeling + interoperability: workflow engines
3. tools to augment existing codebases or scaffolding for greenfield models
4. Value proposition for funders is that we would be improving the reach and longevity of models that are OMF standards compliant
5. metadata validator: establish a schema and build a validator that meet metadata requirements
6. closely support the technology needs of the other working groups and directorate

identify funded model integration projects and work closely with them (via our own internal funding possibly) to get their software to be OMF standards compliant and meet best practices. Once we have a library of compelling examples we concurrently build tools to support that process

the SGCI model of developer-for-hire - being strict on timelines, resources for hire, champions for specific types of things

plugging in models to build a workflow: https://www.esipfed.org/lab/geoweaver

ways to support metadata creation: examples from NASA look to how to create metadata for a dataset: automated by NASA GCMD tools for metadata creation and editing

[cpritcha]

Metadata bots for accessibility checking

• inspirational metadata creation and editing https://uat.earthdata.nasa.gov/data/find-data/gcmd

[franklinzhanggis]

Here may be a paradigm for model meta data http://geomodeling.njnu.edu.cn/modelItem/3f6857ba-c2d2-4e27-b220-6e5367803a12

[alee]

Breakout Room 1

curation services for longevity are great: debugging, removing errors, making models cleaner, more efficient

collect list of exemplars: check with takuya and tony jakeman

look at best practices from https://cybergis.illinois.edu/ and other communities (bioinformatics? astrophysics?)

regarding ethics and ethical guidelines: define constraints, limits of interpretation, assumptions and biases (culled from medical community)

pilot projects that demonstrate value: follow up with @mbarker on how to evaluate and assess value - case studies and tracking use and collecting feedback

[AnthonyJakeman]

Takuya Iwanaga has a possible exemplar of an integrated model
https://www.sciencedirect.com/science/article/pii/S2214581819303726

[alee]

Breakout Room 2

@rkoskela :

• standards -> guidelines
• consider providing cloud resources: connect with Annie Burgess at ESIP for connection to Amazon (cloud compute credits for models that go past the earth science)
• don't jump to interoperability just yet, focus on single / simple models first

@hsu000001 : find some friendly pilot projects to start with (check if USGS has any exemplars, otherwise consider LandLab and others)

should be open to commercial support from companies especially if headquarters move to EU (global exposure). EU open science cloud

interest from USGS in click to run notebooks, followup with @hsu000001