Darren Pilgrim:
> Wietse Venema wrote:
> > Wietse Venema:
> >> Darren Pilgrim:
> >>> Wietse Venema wrote:
> >>>> The "domain in a host" approach appears to be more common with web
> >>>> services: 3346 of 3755 domains used the same IP addresses for the
> >>>> domain itself as for web services (http://www.domain). I had
> >>>> expected a larger number here.
> >>>> Number of domains = 3755
> >>>> with A records = 3147 (following CNAME)
> >>>> with www.domain A records = 3346 (following CNAME)
> >>>> with A records == www.domain A records = 2589
> >>> Umm... those two sections don't seem to match. How do you conclude 3346
> >>> have domain A == www.domain A, then state there are only 3147 with
> >>> domain A and 2589 with domain A == www.domain A?
> >
> > I think I have found where the confusion is.
> >
> > 3346 domains have www.domain A records.
> >
> > NOT: 3346 have domain A == www.domain A as you wrote above.
>
> But that's what you said:
>
> "The "domain in a host" approach appears to be more common with web
> services: 3346 of 3755 domains used the same IP addresses for the domain
> itself as for web services (http://www.domain)."
So, that should be 2589 of 3755. It does not change the conclusion
that "domain in a host" is dominant with web but not with email,
in the sample of domains from postfix/securityfocus/freebsd/php
mailing lists over the past 1.5 years.
Wietse
