Re: [dmarc-ietf] cousin domain definition (was Re: Fwd: Eliot's review of the DMARC spec)

[Matt Simerson <[e-mail]>]

On Jul 8, 2013, at 2:17 PM, Steve Jones <[e-mail]> wrote:

> On Mon, Jul 8, 2013 at 2:12 PM, Barry Leiba <[e-mail]> wrote:
> >
> >  <t hangText="Cousin Domain:"> <snip> </t>
> 
> If it's not too late to change the term "cousin domain" for this, I
> suggest finding another term.  "Cousin" implies a legitimate relation,

New Oxford American Dictionary, cousin:
	• a person belonging to the same extended family.
	• a thing related or analogous to another: the new motorbikes are not proving as popular as their four-wheeled cousins.

Having mulled this over, including time wearing my genealogist's hat, I heartily disagree that the term cousin implies any sense of legitimacy. In my experience, it is uncommon for people to distinguish between natural, step, adopted, foster, illegitimate, maternal, paternal, or even to describe the degree of a cousin. Until further inquiry or precision is requested, cousins are usually just cousins. 

> which this isn't.  I would consider, say, "ibm.com" and "lotus.com" to
> be cousin domains.  I might consider "microsoft.com", "hotmail.com",
> and "skype.com" to be cousin domains.

I think most anyone would accept that those are related domains. The term "cousin domain" is already established and already conveys a sense of illegitimacy.  Adding a new term to the mix with a slightly different meaning isn't going to help anyone, even if it is semantically more correct. 

> Things that try to look like they're related, but *aren't*, are what we're talking about here, and
> I don't think of those as cousins.

But a cousin domain *is* related to the target domain, typically by appearance. It's just that as you pointed out earlier, the relationship is not legitimate. 

Matt