Post

Next up, we’re getting the premiere of Blender’s new short: Singularity. I am seeing it the way it needs to be seen: Maximum front row

I saw a demo of this many months ago, half finished. Seeing this now, fully formed, on such a huge screen with great audio is awesome.

Definitely the most comfortable conference seating I have ever sat in

Really great discussions and demos today at GodotCon. I haven’t seen a talk yet, there are so many interesting folks to meet!

Just draw an incredible PVKK demo on an absolutely enormous cinema screen.!

Next @reduz.bsky.social is talking about how to think like a Godot developer.

How is Godot developed? There is a community, and a foundation. Contributions from the community go through the foundation.

The ideal workflow is image 1. A smooth way to work on submissions. But the reality is image 2. Many things get ignored.

Why do submissions get ignored? Volume of submissions is just too high. Answering properly takes a lot of time. Some contributors can’t take a no, no matter how polite. It’s a stressful situation for everyone.

This organically leads to shiny proposals with barely any use, drama discussions, and endlessly open PRs that will never get merged no matter how much they are worked on.

How can you help? Understand this rationale. Think about whether it’s useful now, rather than in the future. Think like a Godot developer.

Introducing an idea: Feature Scoring, to rate a feature to measure if it should be merged or not. Measuring would be on a CRIS score. Complexity, Reach, Impediment, Stakeholder.

Examples of scoring for each point in CRIS.

Example issue:, just for example: VIM in Godot: Complex, few need it.

Example issue: Interfaces. Important, but technically complex.

An example from a feature request by @reduz.bsky.social himself: Action Area. Not important, undemanded.

Example of symbol refactor: Complexity went from 0.1 to 0.9 by a suggestion to make it easier to implement. This would make it score much higher.

Idea: Have a panel that anonymously judges the scores, for each feature. Submitter can appeal if the feel it is incorrect.

Why do this? What would it fix? Have an objective method for reviewing. Help submitters understand why they get accepted or ignored. Give definitively answers to requests. Project staff from harassment from those who can’t take feedback.

I asked: What about feature requests for things that would create new users of Godot that do not yet exist? The reach may be users who aren’t there yet, the impediment is those users are not there yet at all, and the stakeholder cannot exist as those users are not there yet.

by all accounts (including their own at the time) it was more that they set out to do a blueprint-like node system and did not successfully do it because they didn't really understand the task, so they wisely stopped wasting their time

I'm surprised at this idea that if you build a requested feature and no one uses it, you just... remove it. Is talking to users, iterating, refining, a dying art?

Godot is - both by necessity due to a tiny dev team and by choice due to a very disciplined dev process - very rigorous with its features. Blueprints were a large and labor-intensive feature that was not seeing user takeup to justify maintenance, and was seen as better served by a 3rd-party plugin.

(Personally I worry this creates an echo chamber that eventually solidifies into concrete barriers where progress isn’t possible because only what currently exists is allowed. Blender had this same issue until they finally relented and did a UX pass that then skyrocketed their userbase)

I wasn't at the Godot Engine conference presentation, but I can imagine how Godot Engine CRIS tries to focus research and development energy from frontier technologies to the improvement of existing features.

Do you like money? Government money? Here is a talk about getting better at grant writing, to fund your projects, by @baroness.bsky.social

Disclaimer: It depends. There are always people involved, so no advice guarantees a 100% success rate.