Virtually all technormies don't even know what an HTTP or HTTPS is, never mind the fact that there are other protocols available.
Most webservers configure HTTP to point to HTTPS, and HTTPS internally uses HTTP with SSL encryption around it. Gopher has issues of file format carrying along old format issues(80 char long sentences, weird formatters...). Gemini effectively fixes Gopher. Gopher was available in Chrome for a time, though its issues was styling(chrome didn't bother applying any styling so it looked like your normal garbage txt files transferred over HTTPS). Gopher isn't really worth discussion, beyond Gemini implementing SSL certs and these SSL certs being self-signed. The first point this is mandatory, while HTTP doesn't so HTTP is inherently a MITM attack waiting to happen. I remember back in 2012, 66% of sites or something like that still used HTTP. I'm probably wrong, but it's still a flaw. That said, forcing SSL-certs on, increases E-Waste... EVER SO SLIGHTLY, as effecient decryption and encryption via SSL is done on hardware, not in software(usually...) so old systems without the new security hardware struggles. SSL cryptographic hardware was implemented around the turn of the century(1995 for first implementation of SSL). It was computationally expensive, so HTTP and Gopher were around. HTTPS is a bandaid over HTTP. Gopher has no equivalent bandaid, so any time it asks for user input(are there even still Gopherholes that take user input?) this is MITM-vulnerable.
As for obscurity of protocol: your browser implements a lot of protocols you aren't even aware of. There's the obvious ones, http, and https, but then there's also magnet:// for torrents, ws:// for websockets, gopher:// for gopher(I think Lynx still has gopher support but Chrome doesn't since a decade ago). ftp:// for file transfer. file:// for files(this IS its own protocol and your file browser views files this way).
Gemini's big protocol change is SSL-first, and self-signed SSL certificates. I won't get into the latter because I don't know. I think it's an improvement over blindly trusting big cert authorities but idk much. SSL-first adds a computation tax. This should be enough to drop the argument of accessibility to the protocol for technonormies as all these that I listed are inaccessible protocols, but there for other very good reasons. What it needs is suitable technical merits and implementations, but gemini as a protocol offers none, bar those SSL cert points(you can also use self-signed certs for your HTTPS site but self-signed certs are advised against here iirc)
The main appeal of it, is the format that's a reduction of markdown and easier to parse, and it's wide number of usecases. See gemsub as an alternative to RSS. It's prettier and RSS is a really shit garbage XML format. The fact most people use javascript and static site generators or generators in general for RSS and for HTML indicate something is
terribly wrong with HTML and RSS. I wonder if it's due to XML....
A different question could be raised. Why can't browsers take a number of file types interpreted by file extension and render them according to user defaults. I.E. .gmi markdown, regular .md markdown.
fastn as another example. This would completely reduce a lot of browsers down to the following:
- Accurate interpretation and implementation of protocol spec.
- Accurate interpretation and implementation of document rendering spec given user defaults for browser. Hell, this point could(should where possible?) be extracted out, as PDFs get opened in your system's PDF viewer, or emails in your system's email viewer, images in your system's image viewer... so on
Your file browser already does this with local files, and also works on other protocols too(ftp as an example)
The obvious counterpoint here is "what about sandboxing these for security". Well if your format is pure and not a script-format, sandboxing is unnecessary(windows documents sites with javascript). Even modern javascript developers are realising things are better to be done server-side(look at modern nextjs with server-side components and a general move towards server-side computation) and they return a pure format(html with forms and submissions is still a pure text format). As for those formats that need scripting like this, you need sandboxing of some type, hence a browser, hence the unworkable W3C spec that has...
problems...
for useful results and having spez's smug face looming overhead you'd go to Gopher. It's a nice dream.
