Project direction

[0HyperCube]

Project management

PRs

126 not draft PRs doesn't seem conducive to getting or retaining contributors. People are forced to chase and beg on discord if they want reviews or merges. I'd suggest:

• Timon and TrueDoctor have fantastic technical knowledge and are also quite capable of making reasonable UX decisions. Delegating authority to review and merge small and medium PRs. If touch ups are needed to the UX, this can be done as a follow up. Possibly relevant is @cart's 'Scaling Bevy Development' blog.
• You claim that you "have not had any experienced product designers who deeply understand the product, vision, and design philosophy" to delegate to. This project has been around since 2020; it seems you are waiting for Godot. The core team has lots of experience in the technical and product direction.
• @Keavon prioritise reviews. If you still work full time on this, then perhaps dedicating 1h a the start and end of your day to reviews (although I guess you might still be on an odd sleep schedule so it might be the end and start)?

Roadmap

A selection 2026 (of which there are 3 months):

• Stable document format
• Improved imaging model data format
• Physical measurement units
• Local fonts access
• Context menus throughout the editor
• Timeline panel for animation curves
• Text-on-path tool support
• Colour palette panel
• Non destructive shape builder tool
• Simplified main properties panel
• Basic support for PDF export

The stable document format seems to be progressing well and it is great to see this.

I'm not really sure what is involved in "Improved imaging model data format," however since it is seemingly blocking a number of new features such as #4308 I think it would be great to finish. Perhaps it is related to the List<T> nonsense?

There is not a chance that all of these are implemented. I mention this because of your commit on Fri Aug 28 23:01:48 2026 -0700 'Update the roadmap on the website' so you can't really claim this is an outdated assessment.

You state in the mission statement of your website that "a cohesive product vision […] is the strategy that will make that possible". My personal view is that you've not delivered this. As I've said before, focusing on polishing one use case only and then expanding would seem like a great approach. This roadmap just involves randomly prototyping massive features (e.g. whole products dedicated to just animation).

Project direction

Still poor SVG support

Building a feature complete SVG editor is very complex, but I miss SVG features such as arrowheads, markers, filter effects, text on path, etc. It is technically an alternative to Inkscape; it is not currently something I would really use or recommend.

You shut down decently promising efforts to add filter effects citing "reworking the internal graphical data" (which is going on for many months with little noticeable progress) and the fact that the PR wasn't complete. See 'Roadmap' section above.

Graphene

The async execution / iterator design seems like a potentially interesting experiment. However I'm still not sure I fully grasp how the iterator would play out or how polling would work.

The RFC for attributes is still not merged and you have seemingly (?) not even left any comments on it. The approach seems to have rather a lot of graphene compiler complexity. I cannot really assess if any secret progress is made on this; publicly it seems like work has stalled?

It does seem like a lot of work for both of these RFCs seemingly all falls on the shoulders of TrueDoctor. This doesn't really seem ideal.

Table/List/Instances

The UX at the moment for attributes is laughable. You still seemingly cannot construct a list of three colours e.g. vec![Colour::RED, Colour::GREEN] (linked issue closed without resolution). The merge node only produces List<Graphic> and no other List<T>. You cannot use the « Map » node to map one type to another e.g. [1i32,2,3].into_iter().map(|i| i.to_string()). These difficulties find me very frustrated when attempting to use the attributes.

Desktop app

An official tagged release is probably due at this point. The desktop app is stable and has most of the expected features (thanks to Timon). On June 12 you said "we are too near the release to justify committing to a pre-1.0 numbering scheme" which doesn't seem particularly accurate 3 months later. There is in my view no need to block on a magic 1.0 release.

Conclusion

I hold no influence and you probably will disregard this. These are just friendly suggestions to discuss.