Re: #1132: Allowing IP addresses as path-identiites - Poll result

[Frank Ellermann <nobody@xxxxxxxxxxxxxxxxx>]

Charles Lindsey wrote:

> Frank found a nice syntax for <hostname> in RFC 2396

Credits to Wayne Schlitt... ;-)  The prose for the LDH-rule
is in John's 3696, and of course there's a subtlety, 3696
says "at least one ALPHA in a TLD", not "start with ALPHA".

> its successor RFC 3986 has removed that, and refers you
> instead to RFC 1034 plus RFC 1123.

Maybe we need a 3696bis as a BCP, not only "informational".
The relaxed 3696 idea without optional trailing dot is:

 domain   = 1*( label "." ) toplabel     ; at least one dot
 label    = alphanum [ *( alphanum / "-" ) alphanum ]
 toplabel = [ label ] ALPHA [ label ]    ; at least one ALPHA

The stricter and simpler SPF idea copied from 2396 etc. is:

 domain   = 1*( label "." ) toplabel     ; at least one dot
 label    = alphanum [ *( alphanum / "-" ) alphanum ]
 toplabel = ALPHA    [ *( alphanum / "-" ) alphanum ]

Really sad that this detail didn't hit any "IETF last call"
after 2396, at the moment we can only hope that ICANN won't
break both.

> So it would seem that there is nowhere any standard syntax
> for what a domain name looks like, although "everybody"
> (TINE) seems to know what it is.

The (apparent) consensus here to allow "_" in addition to
"-" is already at odds with what "everybody" thinks to know.

> There is also the interesting question whether you are
> allowed two "-"s in succession.

There might be some restrictions for 2ALPHA "--" at the begin
of a label not limited to "xn--".  But I'm too lazy to check
if that's policy or standard or whatever.

> do we want to be able to distinguish <source-identity>s from
> <path-identity>s by the presence of .SEEN/.MATCH/.MISMATCH,
> or somesuch)?

IMHO a <source> is only relevant for !source.MISMATCH! cases.

For !source.MATCH! we offer !! as shorthand, IMO good enough.

Russ' concept (your source.SEEN! would be new, so far I don't
see the point.  Let's say a site got the path claimed-id!...

If site isn't "lazy" the path (modulo POSTED) would be either
site!!claimed-id!...  or  site!source.MISMATCH!claimed-id!...

IIRC Russ said that the site is "lazy", but still wants to add
some diagnostics.  If that's a <path-identity> it could simply
say site!path-id!claimed-id!...

Or even site!!path-id!claimed-id!... (is that really okay ?)

We're only in trouble (syntactically) if a "lazy" site insists
on adding an IP:  site!ip-address!claimed-id!...  would be a
red duck styling itself as green grouse, only in that case a
site!ip-address.SEEN!claimed-id!...  is better (or necessary)

                     Bye, Frank