> wrote on 16 Dec 2005 16:22:29 -0000:
>
>> what is the problem with putting a single computer into a hosting center,
>> name it mycompany.com,
>> and also let it helo as mycompany.com?
>
>It's not considered an FQDN, it's a domain. Depending on how strict the helo
>syntax test is it will
>fail at this stage.
>
>(What about using a client with better quoting behavior? ;-))
>
>Kai
>
>--
>Kai Schätzl, Berlin, Germany
>Get your web at Conactive Internet Services: http://www.conactive.com
>
Kai,
Um, where do you get this notion? A FQDN is an absolute DNS
object reference with no implied parts - i.e. it resolves to the same
value with or without a final trailing period (ala RFCs 1034 and 1035).
I think what you're trying to assert is two things - 1) you are demanding
a fully qualified *host* name which is a subset of FQDNs and 2) you are
refusing to allow a single identifier in DNS refer to both a host and
a domain (and ignoring the historical domains which only consisted of
a TLD - e.g. ".ARPA"), which is certainly allowed. Now RFCs 952 and
1123 define the requirements placed upon host names, but no "single dot"
rule is visible to my eyes. Please quote what cause you to believe your
statement above. Note that RFC952 has not been obsoleted and even uses
the example of the hosts named "SRI-NIC.ARPA" and "SU-TAC.ARPA"; It also
specifically mentions that a domain may have address(es) assigned to it
and in the grammer (not quite a BNF) lists the production:
<domainname> ::= <hname>
i.e. specificly stating there is no distinction at the naming level between
a domain and a host (other distictions do exist and a name may be overloaded
to refer to both or to neither - e.g. when the zone file for b.a contains
an entry for d.c, but no entry for c alone, then c.b.a is a valid name, but
is neither a host or a domain and need not be resolvable).
Leave the FQDN part out and you can try to base an argument on 2821,
but there sections 2.3.4 and 2.3.5 simply and clearly states that "Domain
names are used as names of hosts and of other entities in the domain name
hierarchy." So there seems little left to argue about and "single dot rule",
even a "no dot" case is allowed in RFC2821 - "consists of one or more", though
I don't believe any such hosts still exist.
You can be as strict as you like, but the rule you would be imposing
is your own (which is perfectly legal - local policy is always allowed - i.e.
"your server, your rules"), but what you propose is not anything based on any
RFC or any other standards document of which I am aware.
Paul Shupak
[EMAIL PROTECTED]
