Community fork of carbonyl

[orhun]

Hey all! 👋🏼

As someone who has been around since the early days of carbonyl, I'd like to bring up a discussion that has surfaced many times over the last two years.

I've seen increasing interest in a fork of carbonyl since there wasn't many activity in the repo recently.
Some points to consider:

• What are the use cases of carbonyl today?
• Would carbonyl benefit from continued development? (What are some of the missing features?)
• Would you be interested in joining this development/maintenance effort?
• Would @fathyb support something like this? (hopefully he's still around!)

Please leave a comment with your thoughts if this is something interesting for you.
My personal motivation is I'd like to experiment with @ratatui in carbonyl and simply get moving with open PRs and issues in this repo.

For reference, we have done a similar community fork for tui-rs in the past and it was a success :)

[Managor]

Important things first: fork name. I compiled some names with chatgpt and pruned the list a bit. Pick your favorite.

Functional Groups With “Carb-/Carbon-” in the Name

1. Carbonyl Family
   Carbonyl (C=O)
   Carboxyl (–COOH)
   Carboxylate (–COO⁻)
   Carboxamide (amides: –CONH₂, –CONHR, –CONR₂)
   Carboximidate

2. “Carbamate / Carbonate” Families
   Carbamate (–NHCOO–)
   Carbonate (–OCOO–)
   Bicarbonate (HCO₃⁻)

3. “Carben-” and “Carb-” Reactive Intermediates
   Carbene (:CR₂)
   Carbanion (R–C⁻)
   Carbocation (R–C⁺)
   Carbyne (·C≡R)

[gold-silver-copper]

Carboxyl sounds pretty cool and rusty, oxidate and all that.

Having a full fledged and supported terminal browser would be pretty cool, landing sixel and other image rendering support is +++

[SandmeyerX]

• What are the use cases of carbonyl today?
• Would carbonyl benefit from continued development? (What are some of the missing features?)

As a developer using ratatui to build a news aggregation/reading TUI tool, I support the continued maintenance of carbonyl. It's really cool. My specific use case involves using a terminal browser to fetch text-heavy web pages — various platform news articles, hot trends, etc. — extract their core content, and present it efficiently within a TUI for reading. Nowadays, anti-spider mechanisms on many sites make traditional scraping difficult, and carbonyl’s ability to provide a full browser environment is a key solution to this problem.

I believe carbonyl’s core value lies in “bringing modern browser capabilities to the terminal,” rather than merely simulating rendering.

• Would you be interested in joining this development/maintenance effort?

If there are new maintenance plans for the project, I’m very interested in following :>

A personal perspective ( feel free to ignore ;/ ):

In an age of information overload, browsing pure content — stripped of visual clutter — can actually improve information absorption. A large portion of news, documentation, and technical resource sites can be effectively consumed via the terminal. From a technical philosophy standpoint, this redefines the very essence of a “browser”: it should first and foremost be a tool for acquiring and presenting content, not just a visual rendering engine.

[megapro17]

Why you're discussing this so deep
Just fork it and update, even the same name will work fine

[orhun]

This issue mostly serves the purpose of seeing who's interested and avoiding duplicate issues.

[bentenjamin]

I'd love to see a fork. While I don't use it often, it was extremely useful when I was setting up my nix machine and I was struggling to install graphics drivers. It would be awesome to see sixel support too!

[philocalyst]

I would love to see a fork :)

Cool be interesting to have a Servo branch, as a way to differentiate?

[orhun]

I'm exploring the codebase a bit and playing around with it on a livestream: https://www.youtube.com/watch?v=pxO4IKRMjLY

[philocalyst]

I really think that Servo integration could resolve a lot of the Chromium building headache; and keep things Rust native. If we're willing to give up JS interop, Blitz is genuinely perfect.

Servo has its share of issues, but rendering to terminal shores up a lot of them.

[orhun]

Here is my observation about carbonyl from the livestream:

In the case of a fork, the maintenance burden might come from:

• Building the chromium. As stated in README.md, it might be a painful process. I personally haven't done this yet but planning to do it on a beefy rental server at some point. (*)
• Maintaining the (17) patches for Chromium/Skia/WebRTC. Those patches are required for the Mojom service which is the bridge between the core and runtime. I'm a bit concerned that some of them are outdated or needs updates. This means that we need to work on the Chromium source instead.

The most requested features that we could focus on are:

• Sixel/image support (Consider Sixel Support #11, Added experimental support for DEC Sixel using sixel_bytes. #206)
• Mouse support (Mouse unsupported #57)
• Address bar & tabs support (Any plan to add address bar or tab manager? #8, someone created carbonyl-terminal for that)
• Other issues mentioned in the 0.0.3 milestone

I have also looked into the codebase a bit for a possible @ratatui integration. It seems very possible, we could potentially swap the whole terminal + cell handling with Ratatui & crossterm and integrate the address bar and tab support with it.

---

On a separate note, there seems to be other "terminal browser" projects which are quite promising:

• https://sr.ht/~bptato/chawan
• https://codeberg.org/janantos/brow6el

[philocalyst]

Would prefer Tab support to be behind a flag, as I believe the common use case is one-off browsing.

[philocalyst]

Thoughts on Servo/Blitz @orhun ?

[orhun]

It could be nice I think. My exploration so far was limited to what we have in the repo and how would managing this fork would look like. So I was mainly focused on the current challenges, but I'm definitely open to changing the approach here for the better.

[philocalyst]

Would totally be down to contribute towards Servo, as I really want it to succeed :)