I understand the impulse to simplify things, but there’s an important question that needs to be addressed before doing so. It is the following: why won’t this new, simple protocol accumulate the same complexity as it’s predecessor over time?
Because no one actually uses Gemini other than to write smug posts about how text only is the best.
The Gemini guys actually think about this a lot so they're working to make the protocol the least extensible as possible or not extensible at all. They plan to freeze the spec forever after 1.0
That doesn't stop anything. There is nothing stopping me from writing up a spec that says "Put an annotation like this in your gemini file and with tools supporting this spec addition, you will see an image embedded". For non compliant browsers you will just see plain text and for new extended browsers you will see an inline image.
I agree that this is possible and that the Gemini project might be futile because of this.
Sounds like they want to guarantee this never actually catches on. Pretty sure they will succeed.
One thing: a version number is explicitly not supported by the protocol or standard. There is only one Gemini version now and forever.
There is realistically no version numbers on the modern web. Its a living standard where new features come in whenever. The underlying communications spec is versioned but this has little to do with possible features and only needs to be updated for performance and security.
That's true for HTML, but not HTTP. HTML is organic - here's your document, good luck making it work. HTTP is clearly versioned with significantly different changing features at each revision.
You don't need to touch HTTP to extend features though. Most of what people care about is the HTML/CSS/JS web api which can extend at any time. No user or developer was ever excited about a new version of HTTP which really only helps performance.