Open Source Sustainability/Reports/SAB 2017-03-15

From NumFOCUS Wiki
Jump to: navigation, search

15 March 2017
(public version)

What follows is my update to the NumFOCUS Sustainability Advisory Board for March 2017 along with a summary of our discussion about my update.

Executive Summary

  • I got good information from the survey I ran asking project leads for what they want out of an all-projects meeting.
  • My 1:1 meetings with project leads are in full swing and are going well.
  • I’m contributing my technical expertise to help improve NumFOCUS infrastructure, starting with our web presence and figuring out a CRM.
  • I do not have any blockers at this time nor significant issues on which I need help except possibly for coordinating a response to the NSF’s CI 2030.

All-projects meeting & collaboration survey

With the goal of determining what kind of cross-project collaboration would be most beneficial to project leads, I conducted a survey asking about their needs.

The survey ran from 28 February to 3 March and received responses from 13 of the 18 NumFOCUS member projects (72.2%).

Detailed results of the survey are available here. Key takeaways from the survey include:

  • Project leads would find it valuable to have a quarterly online meeting, conducted over video teleconference and an annual or semiannual in-person meeting.
  • There is interest from some project leads in opening up meeting attendance to a wider audience of all interest project contributors, or perhaps even the general public. There is even greater interest in communicating the results or proceeds of meetings to a wider audience, such as all project contributors and the general public.
  • There is a general consensus that all-projects meetings should facilitater:
    • knowledge share,
    • collaboration towards shared resources, and
    • multi-way communication between and amongst NumFOCUS and member projects.
  • There is a general consensus that the meeting format should support:
    • timely, structured updates from projects and NumFOCUS staff,
    • discussion about relevant topics, and
    • decision-making as needed.
  • Top topics project leads are interested in seeing addressed are:
    • Funding (fundraising, grant writing, etc.)
    • Contributor outreach and on-boarding
    • Governance
    • Infrastructure and shared resources
    • User outreach and on-boarding
    • Building industry relationships
    • Development best practices
    • Metrics and analytics

1:1s with project leads

On 28 February, I began inviting project leads to schedule one (1) hour meetings with me, the goal being to “meet” each other and for me to learn more about each of the projects, particularly about things I can only learn from leads and core contributors.

The agenda for these calls is the same and is intentionally open-ended:

  • What do I need to know about your project that isn’t obvious from public sources?
  • What’s working well? What are your accomplishments?
  • What’s not working so well? What are your challenges?
  • How can NumFOCUS better support you and your project?
  • Top priorities regarding sustainability (short- and long-term)?

So far the response from project leads has been prompt, engaged, and positive. Everyone I’ve talked with so far talks about their project and their work enthusiastically. They are all interested and engaged in the sustainability project.

As of writing this update, the status of these 1:1s is:

  • complete for 8 of 18 projects (44%)
  • scheduled for 7 of 18 projects (39%)
  • scheduling is in progress for 2 of 18 projects (11%)

I expect to have all the 1:1s completed by early April.

So far, some major themes are emerging from these discussions with project leads:

  • Developing a base of contributors who can act as core contributors and project maintainers is a paramount concern to most all projects. Projects do not need more casual contributors. Rather, they are constrained by not having enough experienced contributors who can review and merge pull requests, guide high-level technical discussions, and perform other project stewardship activities that require a good deal of experience. This is true of projects that are relatively young such as nteract as well as more mature projects such as NumPy. Many project leads would like help reviewing and improving their developer on-boarding materials and processes.
  • Another significant concern is that of finding and building a sustainable funding model in order to reduce reliance on unpaid volunteer labor. Very few project leads I spoke to are paid in anyway to work directly on their NumFOCUS projects. When they are, it is usually not their primary mandate and they are reliant entirely on grant-funding, which is resource-intensive to maintain and often counter productive to longer-term work the more mature projects need to accomplish. Some projects are supported by commercial entities via allowing salaried developers to contribute, but this is mostly done in an ad-hoc way which does not provide for long-term stability.
  • Most projects lack an explicit business model for there project and could use support in this area. One notable exception to this is Software Carpentry and Data Carpentry which do have an explicit business model that is performing well for them in the near-term.
  • Common to most all projects is a contributor cycle driven largely by graduate students. That is, most projects find their most prolific, engaged contributors when they are graduate students, sometimes postdocs. This means that many NumFOCUS projects have a contributor cohort cycle. In a lot of ways this works well. In other ways it means that it is harder to find the contributors needed as the project matures and needs contributors with experience greater than the typical graduate student or postdoc. And, of course, the faculty and researchers with this experience are busier and have more things competing for their attention by this point in their career.
  • All project leads want support for increasing their skills and expertise around open source stewardship. They want NumFOCUS to help bring resources and facilitate knowledge sharing amongst member projects. The exact topics that are a priority to projects varies, but includes: fundraising, marketing and outreach, user and contributor on-boarding, event planning, governance, capacity building, and more.


Facilitating a response to the NSF CI 2030 is still on my radar, though I have no update. For those who have put these together before, do you have a sense of how much work they take? I could use some guidance to ensure I’m allocating enough of my time to oversee this property.


I came down with a head cold just before SIAM in Atlanta and had to cancel that trip.

Next week I’ll be in Austin, working out of the NumFOCUS office. I’ll leave from Austin to attend LibrePlanet in Boston, which is the main conference of the Free Software community.

After that, no planned travel until PyData London in May.

NumFOCUS IT infrastructure improvements

One area that’s emerged as being very important to sustainability is improving NumFOCUS’s IT infrastructure. Though not central to my role as Projects Director, I am lending my technical expertise as needed to rest of NumFOCUS staff to improve areas of our infrastructure that directly impact sustainability.

Current projects include: Exploring how to improve our web presence by migrating to WordPress and implementing a cost-effective CRM.

Recognizing both that my primary role is not to manage NumFOCUS’ IT and that improving our infrastructure will positively impact all of our efforts, including sustainability, I’m mindful about how I am allocating time and energy to these projects.

Notes from Sustainability Advisory Board meeting for March 2017


We discussed how to coordinate a response the NSF CI 2030 RFI. More generally, Dan and Todd both provided good insight into the NSF funding structure and approach.

How do projects find NumFOCUS?

Karl asked about how new projects find and join NumFOCUS. Despite not doing a lot of external marketing, we receive 1-2 inquiries (at least) a week. From these inquiries, roughly one project a month applies for fiscal sponsorship. Leah notes we should start explicitly asking projects how they hear about us, to which there is general agreement. We also discussed how the priority for NF right now is to build its capacity to serve existing projects. Once that capacity is in place, recruiting new projects will become more of a priority.

Metrics and analytics

Karl raised the question about project participation metrics (contributors and users) and inquired about the possibility of NumFOCUS providing some kind of support for metrics reporting across NumFOCUS projects. Christie notes is definitely a priority of the Sustainability program and that there is already some groundwork in place. For example, many projects are already collecting metrics of some sort as part of their annual grant reporting. Christie envisions a NumFOCUS contributor dashboard similar to the dashboards from Wikimedia, OpenStackor Eclipse Foundation.

Action Items

  • Christie will start organizing the response effort today (Wednesday, 15 March) by sending a a kick-off email to the NumFOCUS projects list and reviewing the document Matt Turk put together to get us started.
  • Christie will learn more about the NSF and how it works, reaching out to folks on SAB with expertise in that area as needed.