spf-discuss
[Top] [All Lists]

Re: Re: will resistant MTAs be fronted with commercial antispam gateways?

2004-02-11 11:36:38
"Meng" == Meng Weng Wong <mengwong(_at_)dumbo(_dot_)pobox(_dot_)com> 
writes:
    Meng> What do you think of the new Mail::SRS algorithm?

Only skimmed it, but it looks good.

I'm uneasy about the 11-year wrap around of the timestamp, though.
Are we confident that these won't get archived anywhere?  I can't see
any reason why they should end up in list archives and the like, but
if they do then come 2015 spammers might start using old archives.

Most of these addresses will no longer be valid, of course, but a few
will be, and this will probably make SRS (and those hosts running it)
rather unpopular amongst long-time Internet users.

Why not just add three characters instead of two, and avoid the
wraparound problem?  If we're adding 22 characters for the hash, I
don't think anyone's going to object to an extra character in the
timestamp.  Or alternatively use a granularity of 1 week instead of 1
day.

Let's not go creating a year 2015 problem for ourselves...

        -roy


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