Hans Franke wrote:
> Am 20 Aug 2004 13:25 meinte David V. Corbin:
>
>>>>>Hans Wrote.....
>>>>>
>>>>>>http://lowendmac.com/
>>>>>>(note that they are among the few who got the name thing
>>>>>
>>>>>right ... no
>>>>>
>>>>>>'www.')
>
>
>>First I will agree that nearly every site gets it wrong. But that does not
>>imply that this site got it right!
>
>
>>The protocol [http://] and site [...somewhere.tld] are really independent.
>>For many entities ftp.myplace.com and www.myplace.com can be located worlds
>>apart. To further complicate matters an entity may want independent features
>>using the same protocol [internal.mysite.com www.mysite.com].
>
>
>>For MOST sites http://mydomain.tld should map to http://www.mydomain.tld for
>>convience, but I will argue that is is not a requirement to have "done it
>>right".
>
>
> Yep, you're right here. If there are multiple _computers_ with multiple
> _ip-addresses_ involved. Just, to my unserstandig at least 98% of all
> domain names registered are done so for websites, often running with
> several hundrets other on the same machine. So, why do they tell their
> servers only to react onto the subdomain www. ?
Ahem.
First, as referenced by "
http://www.mydomain.tld", "www.mydomain.tld"
is NOT a subdomain, it's the FQDN (Fully Qualified Domain Name) of an
individual host. Even if the hostname www.mydomain.tld is served by a
farm or cluster, as far as client access goes and as far as DNS is
concerned, it's a single entity, and it's a hostname, not a subdomain.
Second, resolving a domain name as a host breaks RFC definitions and
recommendations. A lot of sites do it (including my company, against my
wishes), but it is improper.
Doc
Received on Fri Aug 20 2004 - 13:25:08 BST