Back to Gemini

I have to admit that I wasn't all that impressed with the Gemini protocol when it was first announced. I certainly liked the idea, but I thought both the protocol and the basic 'gemtext' file format would be too restrictive. Of course, I understand why it (kind of) had to be that way: we all know what happens when we give people too much rope. Still, I thought you've got to give folks some rope. Gemini doesn't provide enough rope to hang a hamster, frankly. And that's a image that's going to stick with me :/

Still, the principle of Gemini seemed sound enough. What I expected to happen was that people -- probably very technically-minded people at first -- would discover it, complain about all the same limitations as I did, and then invent something better -- something more flexible and expressive than gemtext, anyway.

This doesn't seem to have happened. While there are newer, non-mainstream Internet protocols than Gemini, none seem to be particularly superior. Spartan scores, in a way, by not demanding TLS encryption. However, it doesn't extend Gemini's authentication strategy, so it's not really suitable for anything that requires a maintained user identity -- we can't even use client certificates. It doesn't change Gemini's document format, either. There are others, but all have their own limitations.

Still, it's surprising what people have been able to do with Gemini, despite its limitations. It's possible to operate a full, modern-looking BBS, for example, and there are a few about. There are more, better search engines than when I last looked, and newer aggregation tools. There was even, at one point, a Gemini interface to Wikipedia (but it seems to be down at present).

There are also new clients, that make reading Gemtext reasonably agreeable. Five years ago, I couldn't find any client for Linux that would render text in variable pitch -- so I had to write my own. I'm pleased to say that my own efforts have been left behind by the likes of Lagrange.

This all leaves me in a quandary. I'm active in other "small web" communities , and I wonder whether these are likely to be more fruitful than Gemini. Uptake seems to be broadly similar -- Kagi's "small web" list now shows about six thousand sites, and I'm told that there are about the same number of running Gemini capsules. There's also the "IndieWeb" movement, which has complementary goals to Gemini and the "small web", if not exactly the same. It's hard to gauge the uptake of the "IndieWeb", since membership is expressed in terms of principles, rather than protocols. As with Gemini, there seems to be a small core of highly enthusiast participants, and a larger number of hangers-on (like me).

In any event, since there seems to be little enthusiasm for yet another alternative protocol, I've decided for now to go all-in with Gemini. I've written a static site generator tool, and I'm gradually building up my capsule. At present, it mostly contains content that I've converted from other sources and, frankly, it's not all that up-to-date. My oldest articles go back nearly twenty years. But I'm hoping to expand it with new, Gemini-only material over the coming weeks.

I'm not claiming that my ramblings are of much interest to many people, but I do my best. At least my posts are written by a human, which is increasingly uncommon in the mainstream Internet.

So I'll give it a go, and see where things lead. If that turns out to be 'nowhere', at least I'll be able to say that I did what little I could.

Published 2026-02-23, updated 2026-02-23

Back to Gemini

Categories