High level open source & project & governance consulting


URSSI could offer high-level consulting to projects, perhaps as part of incubation or perhaps outside of it:

Someone from URSSI works with them to understand the project and its goals, and suggests policies and mechanisms for the project. This would provide specialized expertise that some small projects cannot afford; which might make the project better (and help them make a better case for competitive funding).

If this was not part of the incubation process, it could also be an activity where URSSI is written into the project’s grant to supply this expertise, or could be provided as a service for a fee.



Yes, please!

And it’s necessary to revisit these processes as the project grows and changes. It’s not a one time thing.

Also we, at least Astropy, needs this now. I’ll be sending emails around shortly looking for folks to help advise us on setting up a mgmt structure which involves paying people for their contributions. This is a big change for us both practically and culturally.




We have had good adoption from the HPC community of the Open OnDemand platform. This more widespread adoption has led to two good outcomes:

  1. A growing community of practice of adopters answering each others’ questions on our discourse instance.
  2. A notable uptick in developers creating useful software extensions to the platform.

For software extensions, we support the standard open source GitHub-based workflow of “fork/extend/pull request” as detailed in our contributions statement:

We are discussing internally as whether this is enough or if we should transition to a more formal community-based approach to allow developers to join the project and contribute directly. I know Airavata followed the Apache Foundation process, as an example. We would be interested in knowing the options of different communities (Marlong mentioned Apache, Eclipse, Mozilla Science, Linux Foundation). It would be interesting to know the URSSI community had any input.



Hi Dave,

We talked about this a bit in the Incubators workshop in Feb (see http://urssi.us/blog/2019/02/25/software-incubator-workshop-a-synthesis/ and http://urssi.us/blog/2019/03/28/report-from-urssi-workshop-on-research-software-incubators/).

My recollection of this discussion was that there isn’t really a single answer, and what model is best for a project depends on a bunch of project-specific details.

1 Like