Re: Federated, decentralised communication on the internet (was: domain names, was: hostname)
On Wed 21 Mar 2018 at 09:50:15 (+1100), Ben Finney wrote:
> David Wright <email@example.com> writes:
> > I don't understand why a home user would not be using a smarthost.
> > Perhaps we're talking about a different group of people. Why would a
> > home user want to relay mail rather than submit it to a smarthost?
> First, note that even if you don't know the reason why someone would
> want to run their own mail server on their own connection, that is no
> argument to arbitrarily deny them the ability to do it.
> So, while you're well within your rights to be curious about why, the
> question doesn't demand an answer. Whoever wants to run a mail server on
> their home connection should by default have the right to do so, for any
> reason or no reason, and doesn't need to explain why to anyone.
Nice of you to set up an Aunt Sally to knock down. I'm afraid my post
wasn't a political statement about the freedom of the Internet but a
technical question about the type of person that I only know to call
"home user". This person is the sort who wonders whether to call the
name of their domain "" or, say, home.¹
(That said, thanks for changing the Subject line.)
> As it happens, there are excellent reasons to want to do this. They are
> no less strong now than when doing this was much more common in the
> 1990s and earlier: in order to retain decentralised control, distributed
> throughout the community, of a decentralised and federated communication
And there are plenty of technical reasons why not everyone wants to do
this, and those have been expressed here when the subject of mail
servers comes up. (Don't bother to ask me for references: I can't
The discussion evolved along the way (with Brian) to centre on sending
to a mail server, and whether it would be better for them to send
My reasoning has been that someone wondering whether to call their
host foo or foo.home is hardly likely to be seeking to relay mail.
If they send email directly from their host, it is likely they are
submitting emails through one of the recognised submission methods,
and it may be best for them that they do.
As these submission methods involve authentication (I believe), then
I want someone to explain to me why having a dot is better then not
having a dot in deciding whether a submitter is genuine. And
without the politics.
This issue is hardly likely to affect the freedom of the Internet, so
please don't talk about whether I'm "within my rights" to discuss it.
> The news for the past decade (and more) has given frequent reminder of
> why it's important to wrest control of our communications out of the
> hugh, centralised choke-points that currently reign. That by itself is
> reason enough to support and encourage more people running mail servers
> independent of those entities.
> The person in question may have additioonal reasons, or separate
> reasons. The point is that email is *designed* and *works best* as a
> decentralised, federated system. We should be asking not “why would
> anyone do this?”, but rather “why have we gone so far in relinquishing
> the ability to do this?”.
> And then take active steps to move more toward federated, decentralised
> communication again.
Carry on, London.²
¹This discussion started at about the time (AIUI) that it became
possible to recommend the use of one of ".home", ".corp" or ".mail"
for home users, a topic that has been discussed here many times.
Had I read
at the time I posted
I would have adopted the terminology ".home user" to make it clearer
who I'm talking about, but I was unaware that that resolution had
passed. From the question "What, on a home LAN, is that used for?",
I derived the term "home user" as a shorthand. As for msg00876,
I couldn't have been more correct in writing there:
". causing some discussion here."
²For any youngsters out there, see