Here are my notes from yesterday's in-person meeting with @timtom @jfaurel @Rory and I.
This day was organized to give @Rory the opportunity to meet other community members and engage in more unstructured conversations. This helped build relationships and provided him with additional context. This post highlights some of the most meaningful conversations, but we also discussed several other points more informally.
Morning
The minutes of the Advisory Committee call are published separately in: May 22, 2025 Advisory Committee
Developer Community call
@antonin_d joined the group remotely. Unfortunately, I had to set out for part of this call, and I am not able to provide an update on those conversations.
Goal Post Conversation
This follows up on the conversation regarding goal posts during May 15, 2025 Advisory Committee
Martin presented to the group the structure of the goal post and his intent to make it a community-driven effort. Overall, there is not enough guidance for the community to understand what needs to be done to move one goalpost further. For example, we could
- Provide templates on how to scope a feature,
- Invite the community to document their use case for a feature
- Let individual add their contact details if they want to help scope or define a specific goal post
- Better advertise the goal post section in the contribution guide in the documentation.
The forum is a good place to better document the goalposts at the Needs Scoping and Needs Solution Definition stages by publishing user stories. Then we can coalesce the different use cases for the Core Dev Group to review and help shape the technical direction for implementation.
It is unrealistic to expect the community to manage the goal post page, as it requires a comprehensive understanding of the project's status. Instead, the Project Manager or the Developer in charge of Community Engagement could update the page on a monthly basis, reflecting recent discussions and releases. This can be accomplished during an open community call.
Afternoon
The afternoon was mostly informal conversation and demo between participants. Here is a lose transcript of what we discussed.
NER Extension
Part of the afternoon was dedicated to playing with the GitHub - sunilnatraj/llm-extension: LLM extension for OpenRefine built by @Sunil_Natraj (see also the conversation in Using local ChatGPT-like LLMs in OpenRefine for data wrangling)
We realized that while we have good documentation on how to set up and deploy the extension, the community has not published many use cases and tutorials. Martin shared the recording of the Data Umbrella webinar "Solving NLP Tasks using LLMs" (video, slides) as a good starting point for building use cases.
Accessing Function in Jupyter Notebook from OpenRefine
Thomas demonstrated how to access functions available in Jupyter Notebook through OpenRefine. This helps to work around the limitations of Jython.