Internet DRAFT - draft-klensin-dotless-terminology-harmful
draft-klensin-dotless-terminology-harmful
Network Working Group J. Klensin
Internet-Draft January 23, 2014
Intended status: BCP
Expires: July 27, 2014
"Dotless Domains", Confusion, and DNS Terminology
draft-klensin-dotless-terminology-harmful-00
Abstract
The history of the DNS has included a great deal of confusion about
terminology that has, in turn, led to discussions in which different
parties have used the same words for different things. For example,
"host name" has been used to describe both fully-qualified domain
names with particular properties and the first label component of
such names. While established inconsistent uses may be impossible to
correct, it is in the interest of the community to avoid increasing
the confusion. There have recently been a number of discussions
about "dotless domains" with at least four different definitions used
or implied in different contexts. This document explains those uses
and recommends avoiding the use of the term.
Status and Stream
The "BCP" category has been tentatively suggested for this document
because it takes the position that the use of the term "dotless
domain" is a bad practice that should be actively discouraged by the
IETF. If there is no agreement on that point, or willingness in the
IESG to move it forward, the document will be treated as
Informational and handled in some other way.
On the other hand, while this document clearly interacts with the
terminology in RFC 7085, it does not directly update that document.
Status of this Memo
This Internet-Draft is submitted in full conformance with the
provisions of BCP 78 and BCP 79.
Internet-Drafts are working documents of the Internet Engineering
Task Force (IETF). Note that other groups may also distribute
working documents as Internet-Drafts. The list of current Internet-
Drafts is at http://datatracker.ietf.org/drafts/current/.
Internet-Drafts are draft documents valid for a maximum of six months
and may be updated, replaced, or obsoleted by other documents at any
time. It is inappropriate to use Internet-Drafts as reference
Klensin Expires July 27, 2014 [Page 1]
Internet-Draft Dotless Confusion January 2014
material or to cite them other than as "work in progress."
This Internet-Draft will expire on July 27, 2014.
Copyright Notice
Copyright (c) 2014 IETF Trust and the persons identified as the
document authors. All rights reserved.
This document is subject to BCP 78 and the IETF Trust's Legal
Provisions Relating to IETF Documents
(http://trustee.ietf.org/license-info) in effect on the date of
publication of this document. Please review these documents
carefully, as they describe your rights and restrictions with respect
to this document. Code Components extracted from this document must
include Simplified BSD License text as described in Section 4.e of
the Trust Legal Provisions and are provided without warranty as
described in the Simplified BSD License.
Klensin Expires July 27, 2014 [Page 2]
Internet-Draft Dotless Confusion January 2014
Table of Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . 4
2. The Humpty Dumpty and Queen of Hearts Syndromes . . . . . . . . 5
3. Dotless and the DNS Definition . . . . . . . . . . . . . . . . 5
4. Recommendation . . . . . . . . . . . . . . . . . . . . . . . . 6
5. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . 6
6. IANA Considerations . . . . . . . . . . . . . . . . . . . . . . 6
7. Security Considerations . . . . . . . . . . . . . . . . . . . . 6
8. Informative References . . . . . . . . . . . . . . . . . . . . 7
Appendix A. Alice References . . . . . . . . . . . . . . . . . . . 8
Appendix B. Down the Rabbit Hole . . . . . . . . . . . . . . . . . 8
B.1. What is a Hostname? . . . . . . . . . . . . . . . . . . . . 8
B.2. An Imagined History of "dotless domain" . . . . . . . . . . 8
Author's Address . . . . . . . . . . . . . . . . . . . . . . . . . 9
Klensin Expires July 27, 2014 [Page 3]
Internet-Draft Dotless Confusion January 2014
1. Introduction
The history of the DNS has included a great deal of confusion about
terminology. That confusion has led to discussions in which
different parties have used the same words for different things,
making it hard to reach reasonable agreements or understand
differences. The term "host name" has been a particular source of
confusion and associated problems (see Appendix B.1). While
established inconsistent uses may be impossible to correct, it is in
the interest of the community to avoid increasing the confusion by
adding new terms with multiple and conflicting meanings.
In recent months, there have been a number of discussions of "dotless
domains", apparently starting with several discussions about possible
new gTLDs (generic Top Level Domains) in the ICANN context. Those
discussions have included statements by ICANN's Security and
Stability Advisory Committee (SSAC) [5] (referred to as the "SSAC
report" below) and the Internet Architecture Board (IAB) [4]
(referred to as the "IAB report" below), a survey of the resource
records associated with each TLD [3] (the "Levine-Hoffman study"),
and an analysis of the use of such domains in various protocols [6].
At least four different definitions have been used or assumed in the
various discussions:
1. Naked domain label without any period(s), including missing the
terminating one (SSAC report and probably the IAB report,
although the latter is less clear).
2. Top-level domain name used without subdomains (several instances
around ICANN including some staff comments).
3. Top-level domain containing address records (Levine-Hoffman
study).
4. Top-level domain containing records other than one with an SOA
RRTYPE and records associated with delegation only (a different
variation on the definition used by Hoffman and Levine that would
consider TLDs that contain, e.g., NAPTR, URI, or MX records and
avoid any confusion about "glue").
The first of these assumes that a "dotless domain" may actually be a
label that is subject to completion or search rules to form a fully-
qualified domain name (FQDN) with more than one label. The others
assume that only top-level domains are intended.
This document expresses no opinion about the desirability or
appropriateness of use of DNS entries referred to as "dotless
domains" (however defined). It is only about the terminology and its
Klensin Expires July 27, 2014 [Page 4]
Internet-Draft Dotless Confusion January 2014
use. If such opinions about desirability are wanted, they can be
found in abundance in several of the referenced documents.
2. The Humpty Dumpty and Queen of Hearts Syndromes
Note: Those for whom these metaphorical references are not familiar
may want to consult Appendix A.
Humpty Dumpty [9] is famously quoted as saying "When I use a word, it
means just what I choose it to mean -- neither more nor less". When
there is actually more than one definition, that approach works as
long as the term is carefully defined, each document and context that
uses it clearly identifies which definition it is using, and that
people remember to read the definitions and know which one applies.
At least when the DNS is involved, those conditions are rarely met,
as evidenced by the discussions cited above.
When one reads a discussion about a "dotless domain" (or a "host
name") without a clear and clearly understood binding to a particular
definition -- especially if strong opinions are expressed about
utility or importance -- one is likely to fall into what we might
call the Queen of Hearts [10] variation on the Humpty Dumpty theme,
i.e., "you have to guess what I meant and, if you get it wrong, off
with your head".
Being headless rarely contributes positively to either protocol or
policy discussions. Neither does confusion about contradictory
terminology, even if one is permitted to keep one's head.
3. Dotless and the DNS Definition
Various very informal uses aside, "dotless domain" actually comes
close to being an oxymoron. The basic definition of DNS concepts [2]
specifies that every complete domain name ends in a dot (representing
the root) whether that dot is actually written out or not.
Specifically, it says
"Since a complete domain name ends with the root label, this leads
to a printed form which ends in a dot."
and
"so a multi-label relative name is often one where the trailing
dot has been omitted to save typing."
Put differently, the trailing dot is always there in user-form FQDNs
Klensin Expires July 27, 2014 [Page 5]
Internet-Draft Dotless Confusion January 2014
whether one sees it or not. If there is such a thing as a "dotless
domain", it is a deliberately relative reference. That is not what
usually seems to be intended by the term, although it is a case the
SSAC and IAB reports explicitly discuss.
Depending on the particular use or context, more exact terms might
include "label", "single-label domain name string", "top-level domain
name used alone", "top-level domain that is not delegation-only", and
so on, including variations on that list.
4. Recommendation
Especially in documents that are intended to represent precise
statements of technology, recommendations, or policy, "dotless
domain" should be eliminated from the vocabulary, replacing it with
terms that actually mean something and have precise interpretations,
such at the examples at the end of Section 3. Even in less formal
statements and documents, the use of "dotless domain" should either
be avoided or carefully examined and questioned to be sure that the
author and reader share an understanding about what is intended.
5. Acknowledgements
This document was inspired by a discussion with Spencer Dawkins,
Patrik Faltstrom, Subramanian Moonesamy, and Andrew Sullivan.
Specific comments from Spencer Dawkins were particularly helpful.
6. IANA Considerations
[[Comment.1: RFC Editor: Please remove this section before
publication.]]
This memo includes no requests to or actions for IANA.
7. Security Considerations
Use of sloppy, imprecise, or confusing terminology or terminology
with more than one definition can easily hide issues that lead to
security holes. This document proposes to eliminate one such
specific case and to warn against others.
Klensin Expires July 27, 2014 [Page 6]
Internet-Draft Dotless Confusion January 2014
8. Informative References
[1] Harrenstien, K., Stahl, M., and E. Feinler, "DoD Internet host
table specification", RFC 952, October 1985.
[2] Mockapetris, P., "Domain names - concepts and facilities",
STD 13, RFC 1034, November 1987.
[3] Levine, J. and P. Hoffman, "Top-Level Domains That Are Already
Dotless", RFC 7085, December 2013.
[4] Internet Architecture Board (IAB), "Dotless Domains Considered
Harmful", July 2013, <http://www.iab.org/documents/
correspondence-reports-documents/2013-2/
iab-statement-dotless-domains-considered-harmful/>.
[5] ICANN Security and Stability Advisory Committee, "SSAC Report
on Dotless Domains, SAC053", February 2012,
<http://www.icann.org/en/groups/ssac/documents/sac-053-en.pdf>.
[6] Moonesamy, S., "The case of dotless domains", 2013, <https://
datatracker.ietf.org/doc/draft-moonesamy-dotless-domains/>.
Version -00, dated 2013-07-13, was used in constructing this
discussion.
[7] Wikipedia, "Hostname", Version captured 2013-09-23, 2013.
[8] FreeBSD.org, "FreeBSD Handbook", Version captured 2013-09-23,
2013, <http://www.freebsd.org/doc/en_US.ISO8859-1/books/
handbook/using-bsdinstall.html>.
[9] Carroll, L., "'Humpty Dumpty' in Through the Looking Glass, and
What Alice Found There", Chapter VI, 1871.
Reprinted in The Complete Works of Lewis Carroll 196 (1939), in
Gardiner, M., The Annotated Alice, New York: Clarkson N.
Potter, 1960, and elsewhere.
[10] Carroll, L., "'The Queen's Croquet-Ground' in Alice's
Adventures in Wonderland", Chapter VIII, 1865.
Reprinted in Gardiner, M., The Annotated Alice, New York:
Clarkson N. Potter, 1960, and elsewhere.
Klensin Expires July 27, 2014 [Page 7]
Internet-Draft Dotless Confusion January 2014
Appendix A. Alice References
The characters, references, and metaphors of Section 2 and at least
the title of Appendix B are to a pair of books from nineteenth
century England [10] [9] that are usually described as children's
stories or fantasies. Parts of both have also been extensively
analyzed as social satire and logic exercises. They are fairly well
known culturally in some areas, having been adapted into multiple
motion pictures and other works. If interpreted as children's books
from nearly a century and a half ago, their match to current IETF
(and related) discussions should act as a caution about self-defined,
poorly-defined and localized terminology. If viewed, instead, as
exercises and demonstrations in logic and, for the first reference,
in the nature of names and naming, the caution should be even
stronger.
Appendix B. Down the Rabbit Hole
B.1. What is a Hostname?
"Host name" has been used to describe both fully-qualified domain
names with particular properties such as address records, a practice
that follows the pre-DNS "host table" use of that term [1] and the
first label component of such names (e.g., "foo" in
"foo.example.com").
A current Wikipedia article [7] illustrates all of the confusion
referred to above: the first label of the domain, the complete domain
name, only some domain names, and so on. It is probably consistent
if read carefully enough, but the distinctions and multiple uses are
very subtle.
The "host as first label" model is also used in the configuration
mechanisms of several major operating systems. For example Windows 7
(and most of its predecessors) gives a computer a name "in a domain"
and handles the two separately. By contrast, current versions of
FreeBSD (9.0 and later) consider a "hostname" to be an FQDN (see
Section 2.5.2 of the FreeBSD Handbook [8]).
By contrast, the ISC DHCP Server uses "domain-name" to designate the
containing domain and "host" to designate an unqualified host name to
which the domain names is appended.
B.2. An Imagined History of "dotless domain"
"Dotless domain" is not the first term to enter the Internet's DNS
vocabulary through less formal discussions and then become a problem
Klensin Expires July 27, 2014 [Page 8]
Internet-Draft Dotless Confusion January 2014
due to conflicting uses. In terminology used for top-level domains,
the dubious distinction of being first may belong, not to "dotless
domain" but to "dot-com". The latter actually interacts with the
discussion about all domain names ending in implicit or explicit dots
in Section 3 above. Because of that interaction, the term probably
should have been "com-dot", rather than "dot-com". More important,
once one believes that "dot-foo" is standard and precise terminology
for the name of a top-level domain, "no-dot-foo" or "dotless-foo"
seems natural and it is only a half-step to "dotless domain".
Author's Address
John C Klensin
1770 Massachusetts Ave, Ste 322
Cambridge, MA 02140
USA
Phone: +1 617 245 1457
Email: john-ietf@jck.com
Klensin Expires July 27, 2014 [Page 9]