Since HTTP is now used as the protocol for almost all Internet activities, it's worth questioning its use. The emergence of Gemini is seen as an attempt to dissociate the net, which has gradually grown together around HTTP.

A good protocol

HTTP remains a good protocol, sufficient most of the time for what is required of it. If developers ended up making it do real-time communication, file sending and receiving, caching, high-load streaming... It's because this protocol and its various iterations have the building blocks to do so. You can twist a tool to create a new one, and that's how I see HTTP today.

Alternatives by subject

The developer-initiated merger of protocols with HTTP has made us forget that most protocols are dedicated to certain subjects:

- file exchange, with FTP

- secure backend authentication, with SSH

- sending e-mail, with SMTP

- receiving mail without copying, with POP

- instant communications, with IRC

- send files peer-to-peer, with BitTorrent

A single interface, the web browser

The disadvantages of using these protocols are manifold:

- weaker security layers than with HTTP, since HTTPS has become the absolute standard

- non-federation of identity management within a web 2.0 project

- multiple interfaces and software

Today's web browser is an OS capable of running highly complex virtual machines. Today, all you need a computer for is to open a web browser, be it Firefox, Chrome or Spotify.

Isn't that the problem?

Let's imagine a single super-browser, capable of loading gemini pages, downloading items via FTP, including an IRC client and a Bittorrent client, as well as a mail client and an identity manager based on asymmetric keys - it's really a classic OS.

In my opinion, this is one of the reasons, albeit a very naive one, why specialized protocols have been abandoned in favor of HTTP: everything is in one place, user habits are simplified. I'd still like to see a Firefox release that simplifies the use of these specialized protocols.