.NAME Registrations Not Conforming to .NAME Registration Restrictions

[ Overview - Restrictions and Enforcement - Domain Listing - Summary Statistics - Conclusions & Implications - Motivation ]

Overview

In 2000-2002, the ICANN New TLD Program coordinated the introduction of seven new top-level domains to the Internet's Domain Name System. Among these new TLDs was .NAME, a namespace intended, according to an appendix to its agreements with ICANN, to be used for "personal name registrations" of the form JOHN.DOE.NAME. However, a large number of .NAME domains do not follow the format specified in ICANN's agreement with the .NAME registry. Rather than matching the first and last names of their registrants, or matching their registrants' commonly-used nicknames or pseudonyms, these many domains instead seem to have commercial, humorous, or other intentions inconsistent with the .NAME charter and the .NAME registration agreements that bind all .NAME registrants; to follow naming conventions other than those required by the .NAME registry; or to reflect defensive registrations performed outside .NAME's official Defensive Registration system.

In recent research, I have documented several thousand domains reregistered within .NAME that seem to be inconsistent with the .NAME registration restrictions as embodied in an appendix to .NAME's agreement with ICANN, and as embodied in the eligibility requirements posted by the .NAME registry and accepted by all registrants. A review of these specific registrations as well as their general characteristics may be helpful in understanding the behavior at issue and in evaluating enforcement of registration restrictions in this and other TLDs.

.NAME's Registration Restrictions and Enforcement

Pursuant to ICANN's agreement with the Global Name Registry (GNR, the .NAME registry), registrations in the .NAME registry must abide by certain registration restrictions. In particular, Appendix L requires that registrations follow certain naming conventions, namely that they fit the firstname.lastname.name pattern . Furthermore, the only names to be added to the DNS are so-called "Personal Name Registrations" which are, as defined in that appendix, "a person's legal name, or a name by which the person is commonly known." A clarification to this definition notes that a "commonly known" name includes "a pseudonym used by an author or painter, or a stage name used by a singer or actor."

However, the Global Name Registry and its registrars conduct no verification of the compliance of any particular registration with this restriction. Section 1.(c).(iv) of Appendix L notes:

(iv) Role of Registry Operator. Violations of the Eligibility Requirements or the UDRP will not be enforced directly by or through Registry Operator. Registrants will agree to be bound by the ERDRP and the UDRP in their registration agreements with registrars. Registry Operator will not review, monitor, or otherwise verify that any particular Personal Name Registration was made in compliance with the Eligibility Requirements or the UDRP. Proceedings under the ERDRP and UDRP must be brought in accordance with the policies and procedures set forth in Appendix M.

Staff of the Global Name Registry confirm that they do not conduct verification of compliance with registration restrictions at the time of registration. An email from GNR staff to the author reports that

We have ... declared that GNR is not in a position to be an arbiter of, and that we will not undertake to police, our own namespace.

Instead, the .NAME registry offers two dispute resolution procedures. As in other TLDs, a UDRP offers redress when "domain-name registration is identical or confusingly similar to a trademark or service mark in which the challenger has rights, the registrant has no rights or legitimate interests in respect of the domain-name registration; and the domain-name registration has been registered and is being used in bad faith." In addition, .NAME offers an ERDRP ("Eligibility Requirements Dispute Resolution Policy") for domains alleged not to meet the .NAME eligibility requirements. The Conclusions & Implications section of this document discusses the extent to which these two policies address the nonconforming registrations documented below.

Specific .NAME Registrations Not Conforming to .NAME Eligibility Requirements

In recent testing and archiving, I have prepared a listing of a total of 5850 distinct domains that seem not to comply with .NAME eligibility requirements. Each of these registrations has as its second-level or third-level domain a word, phrase, or other string of characters unlikely to be the true first name, last name, stage name, or pseudonym of an individual. (Of course, some domains may be erroneously included on this list; the author will promptly investigate and, if appropriate, remove any such domains brought to his attention.) Each of the specified names listed below was registered in .NAME on or before May 16, 2002. Since that time, it is likely that other nonconforming registrations have been added, and it is possible that some of these registrations may have been cancelled or otherwise removed.

For each domain, I have attempted to obtain a variety of information including:

• Current title of default web page, or the fact that no web page is currently available
• Registrant name, city, state, postal code, and country
• Date of domain registration by current registrant and identify of sponsoring registrar

Due to the large size of the full listing of results, it is available in sections by letter of purported "last name" of domain name (second-level domain name):

Full listing of results:
A B C D E F G H I J K L M N O P Q R S T U V W X Y Z numbers

Certain keywords were used to generate portions of these listings; the underlying inference is that since, for example, "university" is not a possible first or last name, any domain registration with "university" as the purported first or last name is invalid. The list of keywords is available here, along with the number of domains found to match each keyword in the second-level and third-level identifier and, optionally, along with the specific domains matching each keyword.

Keyword listing with counts of matching nonconforming domain names
Keyword listing with counts and listings of matching nonconforming domain names

Certain domains were flagged for addition to the listing of results because their purported first names matched their purported last names; most individuals have different first and last names, and so such matching is cause for investigation. These domains are listed here:

Domains with identical purported first and last names

Certain domains were flagged for addition to the listing of results due to registration by a registrant also found to register many other domains. These registrants and their respective domains are listed here:

Top nonconforming registrant listing with counts of registered domain names
Top nonconforming registrant listing with counts and listings of registered domain names

The domains listed above were registered via a total of 44 distinct registrars. Registrars varied widely in their rate of nonconforming registrations; only 3.2% of Go Daddy's registrations were found to be nonconforming, while fully 39.7% of Yesnic's registrations were found to be nonconforming. The number and proportion of nonconforming domains registered by each registrar is summarized here:

Counts of nonconforming registrations by registrar
Counts of nonconforming registrations by registrar, with listings of registered domain names

Summary Statistics

The listing above gives a total of 5850 distinct domains likely not to comply with .NAME registration restrictions. As a proportion of all names in the .NAME zone file at the time of testing, the listed nonconforming domains constitute 7.96% of all .NAME registrations. (This quotient omits names clearly intended only for registrar-registry testing purposes, i.e. "testregcon12345.testing.name" and similar names.)

The 5850 domains found likely not to comply with .NAME registration restrictions can be divided into four major categories:

• Personal name usage of the .NAME space inconsistent with the .NAME registration conventions. Some registrants may have failed to understand the firstname.lastname.name convention or may have preferred to register strings related to their personal names that were, however, not of the format specified in the .NAME registry contract with ICANN. For example, some 561 registrations have as their second-level domain the word "family," as in "smith.family.name"; though this use of .NAME is similar to what GNR contemplates in its agreements with ICANN, these registrations fail to comply with registration agreements since "family" is not in fact any individual's last name. Other registrations include the word "and" in their third-level domain, as in "loraineandandrew.mitchell.name"; again, since no individual is named "loraineandandrew," this registration is also not compliant with .NAME registration restrictions.
• Humorous registrations. Certain of the domain names seem intended to be humorous. For example, "the.best.name" and 389 other registrations have as their third-level domain the word "the." These "the" registrations are not the only registrations making use of the common usage of the word "name"; also registered are "my.cool.name," and a total of 71 other domains with "my" as their third-level domain. It is uncertain whether the registrants associated with these domains intend to use them for commercial purposes.
• Defensive commercial and non-commercial registrations. As discussed in the Conclusions section below, some trademark holders may have found it more cost-effective, desirable, or straightforward to register ordinary .NAME domains rather than to register their marks under .NAME's Defensive Registration service. Such defensive registrations were registered both by commercial firms and by other organizations; for example, a representative of Johnson and Johnson registered 205 brand names such as "childrenstylenol.childrenstylenol.name" using a @jnj.com email address; a representative of the Federal Reserve Board, again seemingly acting in an official capacity and using an official email address, registered thirteen domains including "federal.reserve.name" and "central.bank.name"; a representative of the Sharper Image, using a @sharperimage.com email address, registered ten domains including "sharper.image.name" and various Sharper Image products.
• Commercial registrations and possible cybersquatting. Finally, some of the .NAME domains seem in some ways similar to the so-called "cybersquatting" previously observed in other TLDs. For example, a registrant might register "boston.bruins.name" even if he lacked a trademark on that string and even if the Boston Bruins has such a trademark. A variety of famous names are currently registered in .NAME, including "harvard.university.name," "industrial.lightandmagic.name," "pillsbury.doughboy.com," and "touchstone.pictures.name" -- each registered by a registrant with 100 or more .NAME addresses that do not suggest registration by the respective trademark owners. Other commercial registrants may intend not to resell their names but instead to use their .NAME domains as actual commercial web sites. For example, certain registrants have registered generic .NAMEs that suggest the availability of certain products and services; these domains include "cell.phones.name" and "computer.repair.name."
• Other. Certain other domains do not fit within the four categories listed above. For example, some domains offer or suggest the availability of .NAME-related or DNS-related services; examples of such registrations include "buy-a.domain.name" and "business.domain.name." A total of 119 .NAME domains have as their second or third-level domain name the word "domain," while 19 others contain that word as a substring of one or both identifiers.

The overwhelming majority of domains listed above are not currently providing actual web content. 3093 of the domains (52%) failed to provide a default web page. Of the others, most provided "under construction" or similar pages, or pages without HTML titles; only 122 domains (2.05%) were found to offer functioning default web pages with HTML titles not immediately suggesting that domains were not yet configured, under construction, or available for purchase.

With separately-obtained data about registrar market share among .NAME registrations, it is possible to determine which registrars are most overrepresented and which are most underrepresented in the above listing of domains likely not to comply with .NAME registration policies. Registrar Hangang Systems is found to be most overrepresented among these likely-nonconforming registrations -- with only a 0.05% market share among .NAME registrants generally (37 domains registered), but a 0.38% share (22 domains) among the nonconforming registrations documented here. Among large registrars, YesNic has the highest proportion of nonconforming registrations (417 of its 1051 .NAMEs), while Register.com and Verisign have the largest absolute number of nonconforming registrations (914 and 913, respectively). Among large registrars, Go Daddy has a notably low proportion of nonconforming registrations (213 of 6757, or 3.18%), as do NameScout, Melbourne IT, Verisign, and Register.com.

Certain registrants have registered large numbers of nonconforming .NAME domains. The top five nonconforming registrants have registered a total of 1316 nonconforming names, while the top fifty nonconforming registrants have collectively registered a total of 2500 nonconforming names. Many of those who register domains found here to be nonconforming also register large numbers of domains that current methods cannot flag as nonconforming. However, when a single registrant registers the names of some dozens or hundreds of famous individuals, as did registrants Pascal Leemann-Pluot and Steve Finberg (among others), the most likely inference is that the registrant is not commonly known by any of those names and therefore cannot register them, consistent with .NAME registration requirements.

Of the listed domains, 782 use identical second-level and third-level domain names. Examples include "abc.abc.name," "porn.porn.name," "realtor.realtor.name," and "rebecca.rebecca.name."

Conclusions, Future Work, and Policy Implications

The preceding analysis suggests that at least 7.96% of .NAME registrations (5850 of 73698 registrations) may fail to comply with .NAME registration restrictions. Some members of the Internet community may consider such non-compliance unimportant; some may find the registration and non-commercial use of, for example, "the.funniest.name" to be perfectly acceptable even if formally inconsistent with .NAME's stated registration practices. Nonetheless, other members of the Internet community may seek that .NAME's registration restrictions be followed and enforced -- for purposes perhaps including the creation and maintenance of a particular identity within the .NAME space, compliance with ICANN-registry-registrar-registrant contracts, and satisfaction of the terms of GNR's initial proposal to ICANN. In this context, it may be helpful to review the possible policy changes likely to reduce or prevent the registration of nonconforming domain names:

• Improvements to communication between registrars and customers. When a customer requests a .NAME domain, a registrar might be required to explain to that customer -- in plain and simple wording -- certain major restrictions on .NAME registrations. Such wording might include text along the lines of "You may register only your personal first and last name or a nickname or pseudonym by which you are commonly known. You may be required to produce documentation confirming your name. Violation of registration policies may lead to forfeiture of your registered domain without refund of registration costs."
• Automated systems to prevent obviously-nonconforming registrations. The many domains listed above reflected the usage of a relatively modest list of keywords and other techniques for finding nonconforming registrations. While some of these keywords were generated in response to a search for specific nonconforming registrations, others were predictable nonconforming words or phrases likely to be sought out by registrants. Straightforward software can prevent or discourage such registrations. For example, whenever a customer requests a .NAME domain with first name of "the," a registrar's software might seek to confirm the registrant's intentions, might offer a link to a FAQ describing registration restrictions, and might specifically warn the customer of the risk of cancellation of a domain found to be non-compliant with registration restrictions. Similar prompts, notices, and confirmations might be displayed whenever a customer requests a .NAME domain with matching or including a variety of other strings, or with an identical purported first name and last name.
• Human review. When combined with the relatively simple requirements of .NAME's registration restrictions and the relative ease with which at least a subset of nonconforming registrations can be detected, the size and current rate of growth of the .NAME registry make possible manual human review of each domain as it is registered. A trained staff person could likely review at least several domains per minute; a simple checkbox might be used to flag those names suspected of being nonconforming and to trigger further investigation. In a fully-implemented system, the checking of such a box would trigger the automatic transmission of an email message to the registrant of record, requesting a fax copy of a driver's license or other documents to establish compliance with .NAME's registration restrictions. Such human review might be implemented in a way that improves performance over time; for example, domains with first names and last names previously found (in different combinations) to be acceptable might subsequently receive lower priority for human review. To improve efficiency in the short run, automated systems could be used to search for and identify registrations likely of greatest concern.
• Limits on the number of .NAME registrations per individual. Since .NAME policies require domains to use strings that precisely reflect the name or names by which a person is typically known, and since most people are known by only a few names, .NAME registrations consistent with .NAME restrictions could be limited to several per person. (Registration of more .NAME names could require special administrative approval.) The benefits of this approach are supported by the many registrations by certain nonconforming registrants, i.e. the registration of 200+ famous individuals' names by each of Pascal Leemann-Pluot and Steve Finberg. However, such an approach might encourage interested registrants to adopt multiple usernames, pseudonyms, and identities -- ultimately hindering enforcement and tracking efforts.

The enforcement of .NAME registration restrictions raises questions about registry and registrar incentives. The enforcement of such restrictions would have an unambiguous initial effect of reducing registration counts. In this context, it is not surprising that registrars have hesitated to fully inform customers of the restrictions at issue or to discourage nonconforming registrations; registrars likely hesitate to turn away customers, especially when competitors may have more lax interpretations or presentations of registration rules. Similarly, GNR's direct financial interest in selling .NAME addresses discourages GNR from performing or requiring increased vigilance in correctness of domain registrations. Thus, effective enforcement of the .NAME registration requirements is likely to require centralized coordination of the requirements and policies at issue, i.e. a statement from ICANN about what is expected from registrars and the .NAME registry in this regard.

In some respects, .NAME registration restrictions could be improved in specificity. For example, while registration restrictions allow "[name by which you are commonly known]," current registration restrictions do not speak to certain likely registration requests. Among other areas of uncertainty, .NAME's restrictions do not speak to the question of personal titles; if President Bush wanted to register "president.bush.name" or "president.georgewbush.name," it is unclear whether such registrations would in fact comply with .NAME's registration restrictions. Further clarification of policies is necessary to assure consistency in any future rigorous enforcement efforts.

As documented above, some commercial registrations may have been performed by trademark holders attempting to defend squatting on their names. .NAME registration policies explicitly prohibit this practice; instead, trademark owners are asked to purchase Defensive Registration subscriptions that block registrations of particular strings in second and/or third-level domain names. However, these registrations cost as much as $1500 up-front, while ordinary .NAME registrations cost as little as $20. (Source: Dotster) Thus, seeing an absence of enforcement of registration restrictions, at least some trademark holders may have decided that purchasing ordinary .NAME registrations was more cost-effective than specialized Defensive Registrations. Other trademark holders may have been confused as to the proper procedure for protecting their marks; here again, further work by registrars may be helpful in assuring compliance with registration restrictions. In the future, use of defensive registration services may be more widespread -- and subsequent problems reduced -- if this service is offered at a lower price.

To address the problem of nonconforming domain registrations, .NAME offers two dispute resolution procedures. As in other TLDs, a UDRP offers redress when "domain-name registration is identical or confusingly similar to a trademark or service mark in which the challenger has rights, the registrant has no rights or legitimate interests in respect of the domain-name registration; and the domain-name registration has been registered and is being used in bad faith." In addition, .NAME's ERDRP ("Eligibility Requirements Dispute Resolution Policy") can be invoked against domains alleged not to meet the .NAME eligibility requirements. However, the cost of an ERDRP challenge is at least $1,150 (from NAF) or $1500 (from WIPO). While these prices might be within reach of trademark holders seeking to protect their marks and within reach of celebrities and other individuals determined to hold the .NAME domains referring to their own names, it is unlikely that anyone would invoke such expensive procedures to challenge, for example, the various .family.name and "humorous" registrations listed above. In this context, then, the .NAME registry's current enforcement and dispute resolution policies create a de facto endorsement of the various registrations nonconforming under registration agreements; though .NAME staff write in email to the author, "If there are individuals who wish to dispute the 'family' names, then we would encourage such actions," the cost of such challenges assures that virtually no such challenges will take place.

Comments from the interested readers are welcome and may be linked from this section.

• Adrian Paul Miles, registrant of 220 non-conforming .NAME domains identified above as well as twenty additional .NAME domains, has posted A Reply to .NAME Registrations Not Conforming to .NAME Registration Restrictions. (June 2, 2002)

Update: GNR has submitted a reply to my report to ICANNwatch. (June 8, 2002)

Motivation

The purpose of this work is primarily academic -- to document the activity at issue for the benefit of those who seek to make policy decisions on related matters. For example, ICANN is evaluating and will continue to evaluate the rollout and effectiveness of its New TLD Program; this report suggests certain possible changes to the enforcement of registration restrictions, and ICANN might seek to consider such changes in possible future new TLDs.

This page is made available to inform discussion about the registration of Internet domain names. The data contained here is not intended for use for other purposes, and it should not be used for other purposes without first contacting the author.

Reporting omits seven domains registered by one commercial registrant that had used ordinary .NAME registrations in place of Defensive Registrations but that certifies to the author that it has since canceled its .NAME registrations.