spf-discuss
[Top] [All Lists]

Re: Re: [IETF] Allocation of the new RR type for SPF

2004-11-23 22:09:57
--Alex van den Bogaerdt <alex(_at_)ergens(_dot_)op(_dot_)het(_dot_)net> wrote:

On Tue, Nov 23, 2004 at 07:26:30PM -0500, terry(_at_)ashtonwoodshomes(_dot_)com
wrote:
>> wayne wrote:
>> Personally, I don't think that exclusively using a DNS RR in
>> the spec that won't exist on most people's name servers for
>> many months or years is very productive.

And you are certainly correct for the short term, and probably correct
for at least a couple years to come.

What would be a realistic, conversative, number ?

If we want to get rid of TXT records there MUST be a MUST NOT USE
clause in the RFC.  For instance (example numbers, change to fit):

The following is surrounded by the necessary "spf is optional, if
you are going to implement it then these are the requirements"


I am actually OK with not having a schedule. I would like to have a schedule but I don't believe any of us have the ability to write one that would be at all believeable. I think the only harmful side effect of not having a schedule is a need to support TXT for "a good long time"... probably until there are so few left that nobody cares, or another RFC comes along and declares TXT v=spf1 records as deprecated.

I don't think there is really a rush to get off of TXT records. They don't seem to have an assigned, intended use that we're conflicting with. Mostly I think the reason we want to start thinking about SPF RR type is that we'll all feel a bit silly 5 years from now if we never started down the path at all.

Is there another bad side effect of not setting a schedule that I didn't pick up on? I wasn't sure why you said MUST have it - my guess is that it hinges on that "if we ever want to get rid of TXT" clause...

Thanks

--
Greg Connor <gconnor(_at_)nekodojo(_dot_)org>