On Fri, Feb 27, 2004 at 09:19:08PM +0000, David Woodhouse wrote:
|
| It seems that SPF is a half-way house between the two. It _appears_ to
| be superficially compatible, and fails to make a clean start with a
| fundamentally saner design than the original all-trusting SMTP. Yet it
| still requires changes at _every_ forwarding site on the Internet,
| damaging its chances of widespread adoption for real.
|
I picked forwarding because it seemed like the sector where the
adoption dynamic would have the most leverage.
Here are some sectors we can divide The Internet into.
sender humans
sender MUAs
sender MTAs
sender ISPs
mailing list providers
web-generated emailers
forwarders
receiver MTAs
receiver MUAs
receiver humans
Each sector can respond to a proposal in three ways: they like it, they
hate it, or it doesn't affect them.
Some sectors are larger than others: the space of sender humans and
receiver humans is much larger than, say, mailing list providers.
I aimed for the design that would have the most number of sectors liking
it, and the least number of sectors hating it.
SPF was the best thing I could come up with. Sender humans and receiver
humans like it, because they stop getting virus bounces and joe-jobs.
The only people who really hate it are the forwarders and web-generated
emailers, because it means change, and everybody hates change.
SPF squeezes forwarders from both sides. If forwarders refuse to play
along, the sender humans and receiver humans at either end will
complain, and eventually they will find a way to work around.
As a receiver, if your alumni email address stops working, you'll just
tell people to use a different one. As a sender, if you find that you
can't reach someone through their forwarding address, you'll just use a
different way to reach them. In neither case will the average end-user
take arms against the new paradigm. Only people who have attached some
kind of sentimental significance to the old paradigm will resist; and it
is well within their rights to, just as John Gilmore operates toad.com
as an open relay. And if a forwarder refuses to do rewriting, it's
really not the end of the world; people will just stop using it.
But the average user only cares about being able to get their message
through, not how it gets there.
So forwarders have to adapt.
If there's a different technology that has a better adoption dynamic,
I'm all ears. It's never too late to find a better way.