On 23/03/18 13:55, Dan Purgert wrote: > Richard Hector wrote: >> >> On 23/03/18 11:31, Dan Purgert wrote: >>> Richard Hector wrote: >>>> On 23/03/18 01:17, Greg Wooledge wrote: >>>>> [...] >>>>> RFC 1594 <https://tools.ietf.org/html/rfc1594>: A Fully Qualified >>>>> Domain Name (FQDN) is a domain name that includes all higher level >>>>> domains relevant to the entity named. If you think of the DNS as a >>>>> tree-structure with each node having its own label, a Fully Qualified= >> >>>>> Domain Name for a specific node would be its label followed by the >>>>> labels of all the other nodes between it and the root of the tree. >>>>> >>>>> For example, for a host, a FQDN would include the string that >>>>> identifies the particular host, plus all domains of which the host is= >> >>>>> a part up to and including the top-level domain (the root domain is >>>>> always null). >>>> >>>> Thanks - Having read that paragraph of the RFC, it doesn't seem to >>>> require any particular number of levels, only that all that exist are >>>> present. >>>> >>>> Richard >>> =20 >>> It requires two "levels" >>> =20 >>> 1. the TLD itself >>> 2. the named host >>> =20 >>> Therefore, "com." (that is, the TLD 'com') is not a valid FQDN. However= >> , >>> "a.com." (that is, the host 'a' on the 'com' TLD) is a valid FQDN. >> >> That's not what I see in the RFC. The DNS only seems to care about a >> tree of nodes with labels, not whether a particular node represents a >> host or a network or something else. So if the node in question has the >> label "com", and "All the other nodes" consist of just the root domain >> (with the null label), that should be sufficient. Even just the null >> label followed by (empty list) should be enough. What you say may well >> be what was intended, but it doesn't seem specific to me. > > You seem to be considering a "node" as representing more than it > actually is -- a node cannot simultaneously be "a named entity" AND ALSO > a grouping. A node is a point in a tree. It has a (possibly null) label, and may or may not represent a host. Fair? > I suppose we can consider a FQDN to be somewhat like a mailing address: > > "country" is the root (though unlike DNS, it's "optional" information - > if not specified "the country of origin" is assumed) > "zipcode" is the TLD > "state" is the domain > "city" is a subdomain > "street address" is another subdomain > "addressed individual" is the entity. I wouldn't necessarily count the addressed individual as part of the address. > This isn't a valid address: > > USA Probably not, unless the USPS or whoever is responsible decides it is, which they could. > Nor is anything up to and including > > 123 Main Street > Anytown, Anystate 90210 I would say that is (as long as it exists), and would assume mail would arrive there. What happens after that is up to whoever clears the mailbox. But obviously an email address requires a local part. > Likewise, "com" is not a FQDN, since it doesn't name a specific entity > that is part of the "com" top-level domain. It currently doesn't. But that's just because there's none of the common 'entity' records. If the admin added an A record or an MX record or something, it would. > On the other hand, "lists.debian.org" is fully qualified for the > specific host "lists" of the "debian" domain of the "org" TLD. However, > "lists.debian" is not a FQDN, as it does not ultimately point to a valid > TLD. Agreed (currently) Richard
Attachment:
signature.asc
Description: OpenPGP digital signature