Re: Regarding "Call for Community Feedback: Retiring IETF FTP Service"

[Keith Moore <[e-mail]>]

On 11/25/20 11:26 AM, Hannes Tschofenig wrote:

> For me the question is whether we need to provide multiple protocols 
> for making our documents available to others. 
> I am sure the 3GPP can fetch our RFCs also via HTTPS in 2020.
>
I find these handwaving arguments unhelpful in the extreme as they fail 
to consider the actual issues with using HTTPS instead of FTP.

Vanilla HTTP provides no reliable way to list the documents in a 
folder.  (like FTP XLST or WebDAV PROPFIND)

Vanilla HTTP also provides no reliable way to distinguish a directory 
from an ordinary file from a symlink, and there are symlinks in the RFC 
repository.

Even if we were to support WebDAV, tor me the question is whether we 
should break a stable protocol that has worked for decades and is very 
widely supported in clients, in favor of something new.

Keith

(I will acknowledge the NAT issue with FTP, though I think that PASV 
mode in FTP should still work.  What I don't know is whether there are 
ALGs or other interception proxies out there that break all FTP even 
when it uses passive mode.   It may be that the best feature of TLS for 
this kind of retrieval of non-sensitive information, is that it thwarts 
most interception proxies.)