Or maybe the underlying philosophy is different enough to warrant its own implementation. For example, the Filestash implementation (which I made) listed in the article is stateless and acts as a gateway that proxies everything to an underlying storage. We don't own the storage, you do, via any of the available connectors (SFTP, FTP, WebDAV, Azure, SMB, IPFS, Sharepoint, Dropbox, ...). You generate S3 keys bound to a specific backend and path, and everything gets proxied through. That's fundamentally different to not fit in the mold of other alternatives that mostly assume they own your storage and as a result can not be made stateless by design. That approach has pro and cons on each side