Hi Team, I am available to help with final tasks to ensure the 3.0 release is solid, so please let me know where I can assist. This email includes some initial suggestions. Feel free to reach out to me over slack too.
Specifically for the developers: I'd like to focus a bit on doucmentation. If there are any examples you feel are currently weak, please point them out and I will add better examples and convert them into tests. Regarding release prep and post-release plans: *1. Release Priorities & Timeline* - We can easily follow up with a 3.0.1 or 3.1 release shortly after, so there is no rush on the items below. - Personally, I consider none of these tasks finalized until our core release duties are completed. *2. Blog* - As discussed, where and how should we host the blog? I want to release something on a regular cadence. I'd suggest that outside articles from volunteers will take precedence over those from non-volunteers. *3. Add-ons Strategy* Rather than risking a flood of PRs, I propose using add-ons to engage community volunteers: - *Start with a github issue.* Instead of keeping deferred features in main PRs, I will turn them into issues for volunteers to take on. - *Process*: Anyone wanting a new add-on would create an issue in OpenNLP issues. - *Starter Branches*: If code was previously started for a feature, volunteers can use it as a baseline. (We've seen great results with this recently, where volunteers submitted higher-quality code than the initial draft). - *Lower Barrier to Entry*: Add-ons would have a lower contribution bar to allow for experimentation and specific dependencies (e.g., an S3 outputter). - *Maintenance*: If an add-on proves unuseful, incurs high maintenance overhead, or is abandoned, we can simply remove it. -* Current Status*: I have one add-on ready to release right after the main launch, along with about a dozen tickets for advanced NLP features. I don't plan to push them all at once, rather at the team's pace. Please let me know if you have any thoughts or objections to this approach. Best rungards, Kristian Rickert
