Never implemented Celestia features list for future vibe coding

[john71]

Celestia can be and will be vibe coded in the near future.

It means that Celestia development can make gigantic steps in the next few years.

Please list all planned scientific, visual or other upgrades here which were never implemented because of complexity or lack of coding participants.

My quick list (sorry if I listed features already in Celestia ):

* Real n-body gravity simulation
* Gravitational deflection of spacecraft and asteroids
* Removal of shadow-casting size limits for small bodies
* Shadows for irregularly shaped bodies
* Multispectral / wavelength-based rendering
* False-color astrophysical visualization modes
* Camera optics effects (lens flare, glare, bloom)
* Higher-resolution planetary terrain and texture streaming
* Volumetric atmospheres and realistic cloud scattering
* Procedural galaxies and nebulae
* Interactive spacecraft flight simulation (not just camera navigation)
* Removal of hard-coded time range limits (e.g., ±9999 years)
* Improved and updated spacecraft and station models
* Full galaxy-scale rendering engine (celestia.Sci successor)
* Gravitational lensing visualization
* Relativistic travel / time dilation modes
* VR support (official upstream, not forks)

[Sirius_Alpha]

Perhaps you should show us an example of a vibe coded -improved Celestia. It's open-source software, after all.

[john71]

As I wrote elsewhere "In 2027 anybody with a serious coding AI account can create any modified Celestia binary.". https://celestiaproject.space/forum/viewtopic.php?f=4&t=24487

I wrote this here: "It means that Celestia development can make gigantic steps in the next few years."

I'm not in any form belittling or degrading the fantastic work the current team has done.

They saved this unique project.

This team and forum created many hours of excellent fun for a lot of people and we should appreciate that.

But to be honest, they need all the help they can get.

Isn't it true?

I think sometimes simply by being present and thinking about the future can help.

By 2027 the combination of more capable agentic coding and more affordable token prices could make it feasible for some people to attempt a vibe‑coded version of Celestia.

[Anthony_B_Russo10]

Honestly, considering how old parts of Celestia's code base is, I would not trust AI with it.

[john71]

I think agentic coding can help Celestia most in understanding and modernizing its legacy codebase, not replacing human maintainers. The safest first step is AI-generated documentation, architecture maps, and code analysis with zero functional changes. Next, AI should generate regression and fuzz tests to make future changes safer. AI can then prototype performance upgrades. It can also accelerate new features like procedural planets or VR support. Human maintainers must always review scientific logic, rendering math and data format changes. Legacy code is actually where AI brings the most value, because humans struggle to understand it at scale. IMHO the biggest risk is not AI, it’s stagnation due to limited volunteer time, which agentic coding can dramatically reduce.