4 ms·

That’s a powerful set of scare quotes around “legacy”. FTP, by any reasonable definition, is legacy, and if you need FTP, you have FileZilla, or the terminal.

Spivak 6y ago
FTP is legacy in the sense that it hasn’t hasn’t had the masses giving it love but it’s not dead by any stretch. ProFTPD, pure-ftpd, and vsftpd are all active projects that are shipped with every major Linux distro. It has first class support on Windows through IIS. Every platform has native clients for it. FTP is guilty of not being cool anymore but letting it die completely seems like such a waste. FTP could be the already widely-supported solution to the send a file between two devices problem that has a million bespoke solutions. Oh well.
- viraptor 6y ago
  > FTP could be the already widely-supported solution to the send a file between two devices problem It really couldn't. It has no discovery, no active identity validation story, issues with NAT traversal, relies on filesystem abstraction, ... In sending a file between two devices the actual file sending is the most trivial part and the only part that FTP can do (in the old and clunky let's-split-control-and-data way). We deserve better.
  - Spivak 6y ago
    HTTP also relies on a the filesystem abstraction, it doesn't stop web servers from presenting virtual versions. The advantage that FTP has is ubiquitous and broad device support. FTP can provide the transport and authentication for a huge variety of devices and provide an instant user story of how to connect with older devices. Discovery and NAT traversal are external to the transport protocol anyway and would be just supporting _ftp._tcp records and the normal TURN/STUN servers.
    - viraptor 6y ago
      HTTP does not use filesystem abstraction. The path in HTTP requests is completely unrelated to files and just looks like a file path. FTP has actual stateful filesystem traversal commands. PWD / CWD / CDUP / ... . Of course it can be faked, but why bother? I understood devices in the comment as personal devices for end users rather than servers. In that case discovery is really not external. Same with NAT - it's a difference between "kind of works sometimes" and "is designed to work with it". See Skype's initial success. Or how SIP documents NAT behaviour extensively https://tools.ietf.org/html/rfc6314 https://tools.ietf.org/html/rfc6314
- perardi 6y ago
  Uh-huh. So, uh, use those clients? Chrome will presumably even send ftp:// links to those clients.

Spivak 6y ago

FTP is legacy in the sense that it hasn’t hasn’t had the masses giving it love but it’s not dead by any stretch. ProFTPD, pure-ftpd, and vsftpd are all active projects that are shipped with every major Linux distro. It has first class support on Windows through IIS. Every platform has native clients for it. FTP is guilty of not being cool anymore but letting it die completely seems like such a waste. FTP could be the already widely-supported solution to the send a file between two devices problem that has a million bespoke solutions. Oh well.

viraptor 6y ago
> FTP could be the already widely-supported solution to the send a file between two devices problem It really couldn't. It has no discovery, no active identity validation story, issues with NAT traversal, relies on filesystem abstraction, ... In sending a file between two devices the actual file sending is the most trivial part and the only part that FTP can do (in the old and clunky let's-split-control-and-data way). We deserve better.
- Spivak 6y ago
  HTTP also relies on a the filesystem abstraction, it doesn't stop web servers from presenting virtual versions. The advantage that FTP has is ubiquitous and broad device support. FTP can provide the transport and authentication for a huge variety of devices and provide an instant user story of how to connect with older devices. Discovery and NAT traversal are external to the transport protocol anyway and would be just supporting _ftp._tcp records and the normal TURN/STUN servers.
  - viraptor 6y ago
    HTTP does not use filesystem abstraction. The path in HTTP requests is completely unrelated to files and just looks like a file path. FTP has actual stateful filesystem traversal commands. PWD / CWD / CDUP / ... . Of course it can be faked, but why bother? I understood devices in the comment as personal devices for end users rather than servers. In that case discovery is really not external. Same with NAT - it's a difference between "kind of works sometimes" and "is designed to work with it". See Skype's initial success. Or how SIP documents NAT behaviour extensively https://tools.ietf.org/html/rfc6314 https://tools.ietf.org/html/rfc6314
perardi 6y ago
Uh-huh. So, uh, use those clients? Chrome will presumably even send ftp:// links to those clients.