spf-discuss
[Top] [All Lists]

[spf-discuss] Re: Inconsistency in SPF spec re. "domain-spec"

2006-03-25 01:36:24
Julian Mehnle wrote:
 
I think the SPF-specific requirement that TLDs not start with
a digit is purely arbitrary.

Nope, we copied that from existing RfCs (it's not only in 2396)

DNS doesn't "need" trailing dots either, right?

If you're dealing only with FQDNs, yes.  Or with 2821 SMTP.
There are no optional trailing dots in mail addresses or URLs.
 
Not allowing trailing dots is a purely arbitrary restriction,

Bullshit.  I've told you (collective you) more than once about
John's reasons to do it precisely as he did in 2821.  Back in
2004 when I thought it's a 2821 typo and asked him about it.

If you could guarantee that no SPF implementation got the ABNF
as is right, all silently allowing a dot that's not there yet,
you'd be of course free to "fix" it.  If you can't guaranteee
that all existing implementations are incorrect wrt the spec.
as is leave it alone, it it aint broken don't try to "fix" it.

But allowing trailing dots for single-label but not for
multi-label names, that would be even more arbitrary.

The problem is breaking existing implementations, arbitrary or
not is a secondary point.  Lots of historical decisions appear
to be arbitrary, many are, some are even clear errors, but it's
dubious to break working software without compelling reasons.

I think these two restrictions ("TLD may not start with a
digit", "no trailing dots") should be lifted as part of the
last-minute changes.

Unless you dare copy the USEFOR <toplabel> (in that case I'd
be curious what happens next) you can duly report a threatened
appeal to Mr. Hardie.  Shepherds like to know such hazards.


-------
Sender Policy Framework: http://www.openspf.org/
Archives at http://archives.listbox.com/spf-discuss/current/
To unsubscribe, change your address, or temporarily deactivate your 
subscription, 
please go to 
http://v2.listbox.com/member/?listname=spf-discuss(_at_)v2(_dot_)listbox(_dot_)com