ietf-mailsig
[Top] [All Lists]

Re: DKIM: t=SecondsSince1970

2005-07-15 11:34:01

In <42D7E338(_dot_)5040802(_at_)mtcc(_dot_)com> Michael Thomas 
<mike(_at_)mtcc(_dot_)com> writes:

[discussion about the format of times]

I'd say that this is a perfect example of not sweating the
small stuff. Either will work. Both are possible to program.
We made a decision one way. Unless there's actual harm, I
really see no reason to revisit this kind of nit. The downside
of gratuitious spec churn, however, is real.

In my experience with having the SPF I-D go through public reviews
from other working groups, you *will* get complaints about this.  Or,
at least, I did.  SPF has a %{t} macro variable that specifies the
time, in seconds, since 1970.  Personally, I don't like the variable at
all, but I wasn't tasked with creating a new spec, but with trying to
define SPF as it had existed for a year and a half in the wild.

My recommendation:  If you can change it now, do.  It will save you
time in the long run.


If you can't change it, make sure you take into consideration things
like whether the value includes leap seconds or not and to what
accuracy the timestamps must be to.   (For SPF, it was enough to say
"approximate" and not worry about how accurate the value is.)


-wayne


<Prev in Thread] Current Thread [Next in Thread>