On 11/26/20 9:39 AM, Roman Danyliw wrote:
When considering what do if theftp.ietf.org were to be deprecated, the proposal suggested as few things related to handling document pointers and explicitly turning it off (i.e., no service on that port or resolution of that name). A few parties (https://mailarchive.ietf.org/arch/msg/ietf/XFJOyi9a8KsrF4vmBAghb0Zo3mo/,https://mailarchive.ietf.org/arch/msg/ietf/SZBuNoGOiSajpsvurZ9hQuiNINc/,https://mailarchive.ietf.org/arch/msg/ietf/-nZbFI-D1bp-o-LYGJNA_U1k9fQ/) suggested something less dramatic by putting in a pointer-like README file as the sole content. Is your position that you would preferred to instead CNAMEftp.ietf.org toftp.rfc-editor.org instead? I ask because the directory layout fromftp.ietf.org and ftp-rfc.editor.org is NOT the same. My question is which seems better, when you go toftp.iett.org and the directory you expect is not there, but you find a README on where to go; OR you are redirected to another server which has the content, but no hint on where to go (although it might be clear from the presented directories)
Ok, fair. Unfortunately FTP doesn't support a redirect, so CNAME is the best alternative, and FTP servers have been dealing with CNAMEs for decades. But as you point out, it doesn't work well if the two servers don't have pretty much the same content and organization. Next best alternative is probably to run a vestigal ftp server and put something in the welcome message, but then you're still running an FTP server and may as well just keep serving content too.
Keith
