NAME CONFLICTS Last updated on 25 August 1999

Diane Cabell, Fellow
Berkman Center for Internet & Society, Harvard Law School

Table of Contents

I. Technical Overview
Numbers   Names  Domain Name Management    Root Server    Alternative Roots  Top Level Domains

II. DNS Administration
Privatization of Domain Administration  Domain Registration Competition
Uniform Global Dispute Resolution

III. Name Disputes
NSI Domain Dispute Policy  Trademark Basics     Case Law


The practice of using domain names began as a helpful aid for routing data across the Net. During the gentle days of academic research, the simple policy of handing out a name to the first person that asked for it worked wonderfully well. When commerce hit the World Wide Web in 1992, trademark law came barreling right behind and the judicial debris continues to litter cyberspace.  This module will describe how the domain name system actually works and who makes decisions about it, including dispute resolution policies.  Trademark basics will be reviewed as well as important case law generated by the various uses of names on the Net.

Domain name disputes comprise the substantive area that is closest to being subjected to a uniform global regulatory mechanism. The hazards associated with such a new technical world order and the regulatory power embedded in code were once the province of esoteric academia. [Hear audio archive of Lawrence Lessig, New York New Media Association Panel, "The Internet & Public Policy: Whoís in Control? " June 10, 1998.] As the general public begins to rely on the Internet as an important tool for commerce and communication, however, these issues are beginning to infect the mainstream consciousness.


By now, everyone is familiar with the sight of a URL (Uniform Resource Locator), that combination of letters, slashes, dots, and numbers that makes up a web address (such as You know that if you type a URL into your browser, you will (usually) arrive at your desired destination as quickly as your Internet connection allows.

But how do you get there? Understanding the underlying technology and protocols of this everyday activity is essential to fully comprehending the legal issues at stake. We will use a fictional Harvard student as an example.

The Domain Name System (DNS) is the Internet equivalent of a telephone directory. It is the current method used to identify different computers so data packets may be properly routed across the Internet. There are two kinds of identifiers: numbers and names. Names are optional, but numbers are mandatory.


Each host computer on the Internet is assigned a unique identification number called an "Internet Protocol Address" (IP Address). An IP address is like a street address; it is the target for shipping data. Without an IP address, one is invisible in cyberspace. IANA, the Internet Assigned Numbers Authority (the administration of which is being transferred to the new private organization, the Internet Corporation for Assigned Names and Numbers, or ICANN) coordinates the assignment of numbers under contract to the US Department of Commerce. IANA begins by allocating blocks of numbers to Regional IP Registries known as RIRs (ARIN in North America, RIPE in Europe, and APNIC in the Asia/Pacific region). The RIRs then allocate blocks of numbers to large Internet service providers who in turn pass portions down to the smaller Internet service providers who share them with their users. Allocation is not entirely clerical. Although the task is mainly to avoid giving the same number to more than one host, there are issues having to do with keeping numbers within hierarchies so that routing tables (which direct traffic along each Internet node) remain as short and efficient as possible. While domain names may attract the worldís attention, it is control over numbers that really determines access to the Net.

Harvard will assign our web-venturing student, whose name is "Fred," an IP address when he moves into the dorm. IP addresses currently consist of a numerical sequence (IPv4) that can contain as many as twelve numbers in 4 blocks (the segments of an IP address separated by dots "."). IPv4 consists of more than 4 billion addresses but that has already proven insufficient due to the inefficiencies of centrally managed IP address assignment processes. IPv6 has just been approved to expand the number base. Data packets, including e-mail, can be sent by simply using the IP address, e.g., Fred@ Some domain names may be assigned a static (permanent) address while others are assigned dynamically.


It is often reported that names were introduced in order to spare users the difficult burden of remembering long numbers. Indeed, is easier to remember than but that is merely a side-benefit. The actual purpose is to permit greater efficiency when changes are introduced in the records associated with an individual identity (if the local network is reconfigured or the domain holder wishes to move to a different access provider, for example). Postel, J., Reynolds, J., Domain Requirements, 1984, RFC 920.

In the beginning, all traffic was routed to a single database listing all of the owners of host computers with their corresponding IP addresses (a host table). (Read about the history.) Identical copies of the list was copied by thousands of other servers for redundancy so that queries need not bottleneck at one central locationóthis served the distributive defense-oriented goal of making the Internet invulnerable to destruction. Entering all the new names to that single database and distributing fresh copies of the updated table became increasingly burdensome as the number of participants increased dramatically. Furthermore, the larger the list grew, the longer it took nameservers (computers loaded with name resolution software, such as BIND) to read through it to locate a particular address.

In the 1980s, strategic changes were implemented. Most importantly, the list was divided into hierarchical categories (domains) so that (a) each segment could be updated locally without having to revise the entire list (names no longer need to be globally unique, only unique within an organization as telephone numbers need only be unique within an area code, (b) maintenance of the data was distributed among the owners of the data instead of being assigned to a central administrator. Mockapetris, P. Domain Names - Concepts and Facilities, 1987, RFC 1034, Updated-by: Mockapetris, P., DNS Encoding of Network Names and Other Types, 1989, RFC 1101; Mockapetris, P. Domain names - Implementation and Specification, 1987; RFC 1035, Updated-by: RFC 1101 (above).

Domain Name Management
Domains (also called zones) are simply lists (also called registries); the domain name hierarchy consists of lists within lists. Our student, Fred, has been registered by Harvard [registrar of the second-level domain (SLD) of "harvard"] and placed on its internal directory of names (its own zone file) and so he gets the address Fred@, the (fictional) numerical equivalent of When student "Fred" using the computer wants to send mail to student "Mary" who uses the computer, he will type "mail" His computer sends a message to the Harvard nameserver asking for the IP address of The Harvard server sends it to Fredís computer.

If Fred wants to send mail to student Joe using computer, Fred types "mail joe@" Fred's computer sends a message to Harvardís nameserver asking for the IP address of If Harvard's nameserver has just been booted (meaning it has been recently restarted and has lost the copies of all the IP addresses that it already located) Harvard needs to ask the ultimate authority, a root server, for the IP address of ".edu." When Harvard receives the IP address of the .edu nameserver, Harvard then sends a message to .edu asking it for the number of's nameserver. When Harvard receives the address of MITís nameserver, it sends a query to MIT asking for the IP address of The IP address of (currently, Joe) is sent back in return.

Root nameservers have a database which lists the IP addresses of the nameservers for the top level domains (TLD) (.edu, .com, .net, .jp, .fr). The .edu nameservers have a database that lists the IP addresses of the second level domain (SLD) nameservers within the .edu top level domain (, Harvard's nameservers list the IP addresses of specific Harvard computers and any third level nameservers (e.g. & There can be any number of levels. Postel, J. Domain Name System Structure and Delegation, 1994, RFC 1591.

Root Server
The root server system is a set of around a dozen computers, each of which contains an identical copy of the same root file. One of them, [currently the "A" root server operated by NSI under contract to the United States government (USG)] contains the authoritative copy of the TLD zone file. The authoritative copy is the one that all others trust to be the authentic and contain the most complete and most current list of TLDs. The other root servers, which are owned an operated by several independent organizations, simply ask for a copy from A. Whoever manages the authoritative root may enter or delete TLDs. If .com were to be deleted from the root file, then the next time the other servers updated their copies, every single .com domain name would disappear from the Net. People and websites could still be reached using IP addresses, but it is highly likely that general havoc would reign.

Authority over the root file is another possible source of central control on the Internet. It could be used to dictate registration policy across all of the domains or even to regulate Internet content. The easiest way to keep traffic away from an "offensive" web site is to delete the name from the zone file and the easiest way to require zones to police offensive sites is to threaten to delete the entire TLD. Lest one imagine that this power is unbridled, however, one must remember that the authoritative root gets it status only if the servers beneath it use it (that is, "point" to it) to get their directions. The TLD server owners may decide individually where to point, just as the owners of the root servers may decide individually what TLDs to include in their inventory. If the root is not being administered in a trustworthy fashion, or does not offer access to desirable domains, then the remaining root servers or the TLD servers can choose to reconfigure their machines to point to a different root.

There are competing (alternative) roots in existence now, but they are not widely used, either because of a lack of awareness, confidence, or interest or perhaps because the idea of a fragmented, unstable Internet is unwelcome.

Q. What are the boundaries between setting uniform standards and exercising monopoly power?

Alternative Roots
Alternative roots have been eyed warily by the Internet community because they lead to fragmentation of the Net. An alternative root is simply any other root file that does not receive its copy from the authoritative root administrator. Alternative roots have the capability of directing nameservers to TLDs that are not in the authoritative root database. If an independent techie, for example, wanted to set up her own top level domain (.foo, for example), she certainly is free to do so, after all, itís just a digital database. But how does traffic get routed to that TLD?

The decision to route traffic to one root or another is within the control of the user. When Fred establishes access through an ISP (Harvard acts as an ISP to students), and configures the settings, she enters the IP addresses for the nameservers that she wishes to use to locate information on the Net. Any user can choose any nameserver, including those that point to an alternate root, although most use whichever servers the ISP recommends. An alternative root may or may not include the authoritative root data as well as its own unique files. The current authoritative root, however, does not contain any alternative files. As multiple roots develop, there will be multiple networks, each a little different from the other and this is the kind of fragmentation can be seriously destabilizing. One cannot split a TLD among alternative roots, or else it sort of like reaching different companies when you call a phone number depending on which telephone you use.

Alternative roots have the potential of becoming filters of content. A government could easily build its own root containing only its preferred zones, and mandate its use. Pornography could be banished to an alternative root, or one could be devised as a safe haven for children.

In 1996, alternative networks received notoriety when hacker Eugene Kashpureff went beyond merely offering an alternative when he rerouted all web traffic headed to Network Solutionsí InterNIC to his AlterNIC web site (which uses the gTLDs .sex and .med). He acknowledged doing this as a protest against NSIís monopoly in name registrations. Kashpureff fled to Canada, was extradited and pled guilty to wire fraud charges; he received a suspended sentence. (Read more about this: ZDNN Tech News Now, Internet Hacker Pleads Guilty, March 19, 1998.)

Q. Is competition at the root level better or worse for the consumer?

Q. Will alternative root administrators be liable under misappropriation or other theories by diverting traffic to competing zones?

Q. If you were responsible for thousands of domain names, how would you choose which root to use?

Top Level Domains
There are currently about 260 TLDs in the root zone file. About 250 of them are country code zones (ccTLDs) such as .us, .fr and .uk (United States, France, and the United Kingdom, respectively). Uninett, Domain Name Registries Around the World. The ccTLDs are run by local administrators in each country. Postel, J., IANA Policy on Delegated Domains, These administrators were originally experienced local technical personnel (usually academic) appointed by IANA but as the popularity and market value of domain registrations has increased, national governments are beginning to assert independent control over their ccTLD registries. Authority over individual ccTLD administration is therefore an evolving issue. Registrations in the .us domain are handled by a vast number of volunteers who are supervised by the Information Sciences Institute at USC under contract with NSI which is under contract with the US government. Cooper, A.; Postel, J., The US Domain,/1993, RFC 1480.; The US Domain Registration Services.

There are also seven generic domains (gTLDs), operated by various entities under different restrictions. These domains were originally intended for uses categorized by function, such as .gov for government and .edu for education. The US government runs .gov and .mil. The gTLD .edu is limited to 4-year colleges and universities. The most populous gTLDs are .com, .org, and .net, operated by a private contractor, Network Solutions, Inc., (NSI). The distinctions among these last three categories have never been enforced and have become blurred over time. In mid-1999, almost 90% of all domain names registered worldwide had been registered through NSI. This market advantage got its impetus from the early search engines that only indexed these particular domains. It was furthered in 1992 when NSI was granted an exclusive government contract to operate in these domains as both a registrar (accepting applications for domain registrations from the public) and as a registry (entering that registration data into the TLD zone file). (Read the agreement.)

The registry of each TLD currently sets its own policies about name registration and dispute resolution. The technical parameters of name registration (how many characters it may contain, where the dots must appear, etc.) are the subject of protocols (determined by consensus of the engineering community). Beyond that, neither the government nor the technical administrators have heretofore attempted to impose any standards on registration policy; however, this is increasingly likely to change as use of the Internet accelerates.

Although ccTLDs were originally envisioned as being limited to domestic use, the registration policies of the various country code domains have evolved quite independently. Some still restrict themselves to local entities (Italy) but others sell registrations on the open market as if they were gTLDs (like the country of Tuvalu, which owns the coveted .tv domain). Some prohibit registration of product names or trademarks (Norway). Only a few (like Sweden) have a separate SLD for personal use.


NSI ordinarily registers the first applicant to request any particular name, unless there is a complaint from the owner of a pre-existing national or federal trademark. Network Solutionsí Domain Dispute Policy, Revision 03, Effective February 25, 1998. NSIís dispute policies have led to much trademark litigation, as will be discussed below in the section on Domain Name Disputes. While NSI remains a sole source government contractor, however, it enjoys immunity from anti-trust claims in the US, if not abroad. Tech Law Journal, Judge Rules in Domain Name Registration Suit, (April 10, 1998) and see collection of cases at The Perkins Coie Internet Case Digest, Antitrust. Elinor Mills, DOJ Renews Domain Name Registry Probe, Computerworld Online News, (May 6, 1999) (EU/UK anti-competition activity).

Privatization of Domain Administration

In 1996, as the growth of private and commercial use exploded, the Clinton administration recognized that the Internet was now such a broadly distributed network, with ownership of its constituent parts so widely scattered, that it was no longer appropriate for one single government to be coordinating its functions. There was also great dissatisfaction with NSIís exclusive position in the most popular market. Following public hearings (including online sites for public comment), the government decided to withdraw from the role of technical coordinator and permit the private sector to assume this responsibility. US Department of Commerce White Paper, Management of Internet Names and Addresses, Supplementary Information (Background), National Telecommunications and Information Administration, Docket Number: 980212036-8146-02 (June 5, 1998).

In November 1998, the DOC entered into an agreement with ICANN giving the latter one year to develop methods and procedures for technical coordination in a manner that will ensure bottom-up coordination, market competition, global representation and stable operations. Memorandum of Understanding Between the U.S. Department of Commerce and Internet Corporation for Assigned Names and Numbers (25 November 1998). Under that agreement, ICANN assumed the governmentís oversight responsibilities for names, numbers, protocol parameters and the authoritative root and has since undertaken a series of steps to meet its MoU obligations.

Q. By inviting ICANN to assume responsibilities that were previously the responsibility of the Department of Commerce, is the USG transferring public assets or regulatory functions to a private entity?

Q. If so, under what authority does the USG transfer these responsibilities?

Q. How does a regulatory body assess the value of online public comment, given the absence of sender authentication?

Domain Registration Competition
There are two areas in which competition for domain registrations may exist: that of the registrar and that of the registry. The registrar collects a fee and data from the individual applying for a domain name. The registry charges a fee to the registrar to incorporate that data into the zone file. In many TLDs, the two functions are performed by the same entity.

In order to renew its exclusive contract as the com-org-net domain registry, NSI was recently required to admit competition at the registrar level. (Amendment 11) However, since NSI retains exclusive control of the zone files themselves, it remains at a competitive advantage over the other registrars both in terms of price (it charges the them $9 per registration to enter their data) and in terms of access (it enters its own clientsí data as well). Although five testbed registrars were approved by ICANN in April 1999, only two had begun operation by July. There are more than 50 approved registrars waiting for interoperability with NSIís databases.

Registration with most TLDs is automated. (Example) The registration information is entered directly by the applicant on a registrarís website and then automatically compiled in a backend database. Easy access to domain registration data is critical to the stable operation of the Internet. Much technical havoc can be generated that can only be resolved at the source, therefore it is necessary to know who actually operates the domain ownerís nameservers. Plaintiffs also want to know whom to serve in cases of web site malfeasance. This information has always been available through Whois, a searchable archive of every single domain name and IP address registrant in the com-org-net registries. Whois is also used to see whether a particular name is still available for registration. The Whois database, which is now an extremely valuable file, was established by NSI under its USG contract and NSI therefore has control over how much of the personal data is displayed on a Whois search. First Amendment activists want such data suppressed in order to protect anonymous speech. A. Michael Froomkin, Anonymity and Its Enmities, 1995 J. ONLINE L. art. 4. EU privacy rights may also conflict with public accessibility to some of the Whois data.

Q. What is the nature of a domain holderís right in the domain registration? Is it a property right?

Q. Does NSI have a protectable right in the Whois database? Can blank space in a file be owned by anyone? If so, what is the nature of that right? If NSI doesnít own the data, who does?

Q. What right does the USG have to the zone files under its contracts with NSI?

Q. Is there any need for a centrally managed registry if data from all registrars can be entered into the zone file automatically?

Another level of competition is the registry level. One of the most eagerly awaited commercial decisions relative to the root file concerns the addition of new gTLD registries (.biz, .web and .shop are some of the proposals) that would be entirely independent of, fully competitive with, and possibly as lucrative as NSIís. NSI currently maintains the authoritative root file and must cooperate to open this gate, an action that appears to directly conflict with NSIís commercial interests. Q. Is NSI obligated to make additions to the root file at the direction of the USG?

Q. How much can NSI cooperate in the process of admitting competing registrars without violating its obligation to its own shareholders?

There are many benefits to consumers that might result from the addition of more top level domains including longer and more secure registration periods and better prices. It would also ease the perceived shortage of names in cyberspace that exists because there are so many different users who share the same name in real space (such as MacDonald) but only room for one of them at "". Unfortunately, there are also many complications. Unless these new extensions can add significant distinction to a name, they will simply offer more opportunities for trademark infringement. A further unresolved technical question is how to effect domain portability: if a domain holder cannot carry his own name with him from one registry to another, more competitive terms may be of little consequence.

Uniform Global Dispute Resolution

The Memorandum of Understanding with the USG directs ICANN to consider the question of adding more gTLDs. As part of that review, it is requested to take into consideration recommendations made by the World Intellectual Property Organization (WIPO) concerning: (i) the development of a uniform approach to resolving trademark/domain name disputes involving cyberpiracy; (ii) a process for protecting famous trademarks in the generic top level domains; (iii) the effects of adding new gTLDs and related dispute resolution procedures on trademark and intellectual property holders; and recommendations made by other independent organizations concerning trademark/domain name issues.

The Final Report of the WIPO Internet Domain Name Process, presented to ICANN in May 1999, proposes a uniform, mandatory domain name dispute policy (administered by WIPO) for all .com, .org and .net registrants. Final Report of the WIPO Internet Domain Name Process, WIPO Publication No. 92-805-0779-6 (April 30, 1999). Even some of its panel members have expressed concern over some of the recommendations. Prof. A. Michael Froomkin, A Commentary on WIPOís The Management of Internet Names and Addresses: Intellectual Property Issues. If endorsed by ICANN, these procedures are expected to be imposed on domain holders by contract. ICANN took the WIPO recommendations under advisement and referred them to its Domain Name Supporting Organization, but also voted to require all new com-org-net registrars to adopt a uniform domain dispute policy although it left the choice of the terms of such a policy up to the agreement of the registrars themselves. ICANN Berlin Meeting Preliminary Report, Resolution on Report of World Intellectual Property Organization. Those that are currently operational have chosen to adopt NSIís policy. Dispute Policy.

Q. Is the WIPO uniform dispute policy a good solution to the cybersquatting/cyberpiracy problem? Is it a good solution to the domain name/trademark conflict?

Q. If the uniform policy only applies to gTLDs, will those registries be at a serious competitive disadvantage to the ccTLDs which are not bound by the policy?

Q. If the US and other nations pass individual anti-cybersquatting initiatives (see S.1255), how will this impact on any uniform policy directives from ICANN?

The Department of Commerce has given ICANN a one-year probationary period to establish a stable, competitive, representative, and bottom-up system of management for IP address allocation policy, root server oversight, and coordination of protocol parameter assignments). As a private, non-profit corporation established under California law, ICANN has been attacked as unrepresentative because its volunteer 9-member Interim Board was not publicly elected. In response to public criticism and government pressure, ICANN restructured itself to include an at-large membership (equivalent under California law to shareholders) that will elect half of its eventual 18-member Board. Internal and external review systems were also designed based on public input. What was once a lean, mean and very efficient technical administration is being rapidly layered with what some call "checks and balances" and others call "bureaucracy". To many Netizens, it looms as a new international regulatory body that will impose its will by a series of contracts of adhesion. Should there be centralized government in cyberspace? Lawrence Lessig, Reading the Constitution in Cyberspace, 45 Emory L. J. 869 (1996). See archives of the IFWP discussion forum which debates ICANN issues.

Unfortunately, ICANN has not developed a revenue stream at this time and remains an underdefended target of those who oppose any centralized oversight of the Internet. An effort to impose a domain registration fee was attacked in Congress as an unauthorized tax. Hear the audio archive of the July 199 House Commerce Committee hearing on "Domain Name System Privatization: Is ICANN Out of Control. With less than a year remaining to accomplish the task of centrally coordinating the vast multitude of conflicting interests which the Internet has generated, ICANN remains in a very fragile position.


Trademark conflicts have been the primary cause of most Internet litigation to date. This module will not cover disputes over general web site content, as those do not, outside of jurisdictional questions, raise questions that are unique to the Internet. This module will first consider disputes over the right to use a particular word in a domain address and then look at other unique hypertext applications such as links and metatags. Linking, tagging and framing may also present copyright issues. For a detailed discussion of the latter, see the module Copyright in Cyberspace.

There are two ways to resolve disputes between trademarks and domain names. One is through the dispute policy of the domain registrar; the other is through the courts. Each registrar may establish its own dispute policy.

Network Solutionsí Domain Dispute Policy
There are more than 7,500,000 domains registered worldwide as of mid-May 1999. Over 60% (5 million) of them are in the .com registry. The TLDs .org and .net contain another million. (Domain Stats) NSI has a domain name dispute policy that is imposed by contract on registrants. Network Solutions Domain Dispute Policy, Revision 03 (effective 02/25/98). There are many contentious features of this policy, not least of which are the rights reserved by NSI (a) to modify the policy at any time (Sec. 5) and (b) to revoke, suspend, transfer or otherwise modify any registration at any time at its sole discretion (Sec. 7). See Oppedahl & Larson LLP, NSI Flawed Domain Name Policy information page for a review of litigation concerning NSI's dispute policy.

NSIís dispute procedures (Sec. 8 and 9) give preference to owners of national or federal (Principal Register only) trademarks over anyone who holds a domain with an identical name unless the domain was obtained before the effective date of the trademark registration. This is certainly the least expensive enforcement mechanism for those who actually have federal trademark registrations. The mark owner is simply required to send a letter of complaint to the domain holder and then send to NSI a copy of that complaint with evidence of delivery plus a copy of the trademark registration certificate. NSI will then contact the domain holder and ask for evidence of federal mark protection that predates the complaint. If none is provided, then NSI may opt to put the domain "on hold" (prohibiting anyone from using it), evict the domain holder or take any other action that it deems fit, such as transferring the domain registration to the mark owner. If litigation is commenced by either party, NSI will maintain the registration status "as is", unless directed otherwise by a court of competent jurisdiction. Not all domain registrars are so compliant. Oggi Advertising Ltd v. McKenzie and others, CP.147/98 (High Court of New Zealand, Auckland Registry) 2 June 1998. See also Domainz Media Statement, 13 November 1998.

Q. Why not use a simple first-come, first-serve policy for registrations?

Q. Is there an inherent quality to a web address that gives trademark holders a perceived right to control use of their mark (or similarity) when used in one?

By ignoring design marks, common law marks and state registrations, NSIís policy provides no relief to most mark owners. Confusingly similar words and common misspellings, for example, are not addressed because the policy only applies to identical character strings. It offers no protection to fair use and disregards the privileges of non-commercial users. Diane Cabell, The InterNIK Dragon: A Fable for Cyberchildren, The Domain Name Handbook: High Stakes and Strategies in Cyberspace by Ellen Rony and Peter Rony, R&D Books, 1998. Nonetheless, like the proverbial magic bullet, NSI has passed through almost every attempt to hold it responsible for the resulting domain name confusion. Lockheed Martin Corp. v. Network Solutions Inc. , 985 F. Supp. 949 (C.D. Cal. 11/17/97); Dan Goodin, Another Favorable Ruling for NSI, CNET May 12, 1999.

Five new registrars have recently been admitted to compete with NSI to service the com-org-net domain complex. More than 15 more are waiting to come online. ICANN recommended that all of the com-net-org registrars adopt a uniform registration policy of their collective choosing. At present, they have adopted NSIís policy.

Where NSIís dispute policy is of no avail, parties have naturally turned to the courts, which have shown a great deal more respect for the existing principles of trademark law

Trademark Basics
What is the purpose of trademark protection? The Lanham Act, 15 U.S.C. §§ 1051 - 1127, was intended to make "actionable the deceptive and misleading use of marks" and "to protect persons engaged in . . . commerce against unfair competition." 15 U.S.C. § 1127. In essence, the purpose is to protect the consumer from waste and fraud. Important elements of trademarks are that 1) they must be distinctive 15 U.S.C. § 1052, 2) as long as there is distinctiveness, there can be multiple legitimate uses of the same mark, 3) rights in marks relate only to commercial activities and 4) if use is abandoned or if the use loses its distinctiveness, rights lapse.

A trademark is either a word, phrase, symbol or design, or combination of words, phrases, symbols or designs, which identifies and distinguishes the source of the goods or services of one party from those of others. 15 U.S.C. § 1127. A service mark is the same as a trademark except that it identifies and distinguishes the source of a service rather than a product. Marks must be distinctive so the customer can tell which products are made by Sony from those made by General Electric.

Marks are generally classified in categories of increasing distinctiveness: (1) generic (describes the type of product itself, for example, "Ice Cream" ice cream); (2) descriptive (literally describes the product or its attributes such as "Cold & Creamy" ice cream); (3) suggestive (connote qualities about the product without literally describing the qualities, such as "Blizzard" ice cream; (4) arbitrary (unexpected in the context of use, such as "Diesel" ice cream); or (5) fanciful (invoking imaginative images in the context of use, such as "Home Run" ice cream") and (6) coined terms (words that are made up purposefully to be used as a mark, such as "Zerious" ice cream). See Abercrombie & Fitch Co. v. Hunting World, Inc., 537 F. 2d 4, 9 (CA2 1976). The last four categories of marks are deemed inherently distinctive and are entitled to protection. In contrast, generic marks [those that "refe[r] to the genus of which the particular product is a species," Parkí N Fly, Inc. v. Dollar Park and Fly, Inc., 469 U.S. 189, 194 (1985), citing Abercrombie & Fitch Co. v. Hunting World, Inc., 537 F. 2d 4, 9 (CA2 1976)., supra] are not registerable as trademarks. Parkí N Fly, supra, at 194.í For a more comprehensive discussion of these categories, see the Abercrombie opinion. Personal names and surnames (Byrdís ice cream) and geographic terms (Alaska ice cream) are not considered inherently descriptive. Despite these strictures, even generic and descriptive marks may get protection if they acquire "secondary meaning." This happens when the mark user takes the risk of building a brand identity in the absence of strong legal protection and does indeed establish a distinctive name in the mind of the consumer, such as Ben & Jerryís Ice Cream.

Q. If the company CreditComm Services has acquired secondary meaning, will it have an action for infringement against There are two different ways to get trademark protection, either by actual use in commerce under common law or by registration under statute. In the US, the first to use the mark in commerce has priority of ownership and actual use is a condition for registration in most states. If one places the goods/services in interstate commerce (or intends to do so shortly), federal registration is available under the Lanham Act. USPTO (searchable database and online registration forms. In most foreign countries ownership belongs to the first to register the mark. The important point is that the same mark may be legitimately used by different owners in different jurisdictions. This multiplicity of commercial rights (not to mention non-commercial rights to use language for other purposes) conflicts with the DNS identification function which only admits one user to a particular character string.

Because the essence of a trademark is to identify the origin of a good, the owner of a mark must aggressively maintain its distinctiveness or else lose it through the principle of abandonment. The mark owner therefore has the right and obligation to prevent another from using the same or a confusingly similar mark on competing products. "Confusingly similar" is not an easy standard for a layman to follow. It includes foreign language translations of the mark, different words that nonetheless convey the same meaning, and variations in spelling.

To determine whether there is a likelihood of confusion, courts look at a number of factors including: (1) the degree of similarity between the marks in appearance and suggestion; (2) the similarity of products or services for which the name is used; (3) the area and manner of concurrent use; (4) the degree of care likely to be exercised by consumers; (5) the strength of the complainantís mark; (6) actual confusion; and (7) an intent on the part of the alleged infringer to palm-off his products as those of another. Some courts add additional factors, such as whether the senior mark holder is likely to "bridge the gap" and start marketing its services or goods in the markets in which the junior user uses the mark. David Loundy, Primer on Trademarks.

Q. Should the Internet be considered a single area of use for assessing likelihood of confusion?

Q. If you believe that it should, how would you assess a conflict between (the website for a small pie place in Brooklyn), and Tony's Pizza (a national chain of pizza makers)? How would the other factors bear on your assessment?

Q. If you believe that it should not, how do you respond to the assertion that the "area of use" factor was designed for the days when local businesses had no possible of conflicting with the branding of businesses in far-off locations and all that has changed with Internet use?

There is also an additional statutory protection for famous marks. "Anti-dilution" statutes exist in many state and national jurisdictions to prevent impairing the value of a famous or well-known trademark. [See,e.g., 15 U.S.C. 1125 ] The Federal Anti-dilution Act was specifically intended to apply to domain name. (See Loundy, supra, at fn 32) The Federal Trademark Dilution Act defines dilution as "the lessening of the capacity of a famous mark to identify and distinguish goods or services, regardless of the presence or absence of: (1) competition between the owner of the famous mark and other parties; or (2) likelihood of confusion, mistake or deception." This statute provides for claims against non-competing and non-confusing uses, which have the effect of diminishing or diluting the effectiveness or distinctiveness of a mark. Under anti-dilution, the mark owner may prevent any new commercial use of the mark, regardless of the lack of competition or confusion if the new use blurs or tarnishes the value of the famous mark by rendering it common place through overuse or by associating it with unsavory materials. Q. What about generic marks that become famous, for example, Tide laundry detergent? Can it prevent new enterprises, such as a vendor of nautical maps, from using Unlike copyrights or patents, trademark rights can last indefinitely if the owner continues to use the mark to identify its goods or services. The term of a federal trademark registration is 10 years, with 10-year renewal terms.

The US Patent & Trademark office accepts domain names for registration as trademarks if the applicant offers services via the Internet. Further, specimens submitted in support of the application to show use of the mark must show use of the Internet domain name as a source identifier, not as a mere directional reference. In other words, it must be used to label the services, not simply as the equivalent of a telephone number or business address. Also, providing a service which is normally expected or routine in connection with the sale of oneís own goods is not a registerable service. By analogy with the registration of trade names, the more distinctive the presentation of the Internet domain name and the further it is physically removed from other informational data appearing on the specimen, the more likely the name will be perceived to function as a service mark. See the USPTO's Trademark Examination of Domain Names.

Closely related to trademark, and often of greater utility, are rights under principles of unfair competition/unfair trade practices and misappropriation which exist in most states under both common law and statute. Michelle L. Spaulding, The Doctrine of Misappropriation (1998).

Trademark and domain names are a poor fit. One admits many users to a particular space, the other admits only one. Some have suggested substituting a DNS taxonomy that inserts additional identifiers (e.g., and Unfortunately there are multiple legitimate mark owners within such categories and many businesses whose product classification may not be intuitive. Just as important is the fact that Net users avoid long domain names.

Case Law
Courts have generally applied traditional trademark and unfair trade practice analysis to disputes involving domain names.  See the module Jurisdiction over Cyber Actors for more detail.
Jurisdiction & Liability
Although there are some decisions to the contrary (see, e.g., Inset Systems Inc. v. Instruction Set, Inc., 937 F. Supp. 161, (D. Conn. April 17, 1996)), most US courts have held that passive websites that lack interactivity are insufficient to establish personal jurisdiction in a non-resident forum.

In Bensusan Restaurant Corp. v. King, 1996 WL 509716, 937 F. Supp. 295 (S.D.N.Y., Sept. 9, 1996), affíd, 126 F.3d 25, 1997 U.S. App. LEXIS 23742 (2d Cir., Sept. 10, 1997), the court held that uploading a web site onto a server located outside the forum was insufficient, without more, to subject the webmasters to personal jurisdiction where the site clearly indicated that there was no intent to sell products to forum residents. This case involved a local music club that had the same name ("Blue Note") as a well-known New York jazz club. The web site included disclaimers of any association with the New Yorkers and only offered ticket sales by telephone within Missouri.

See Music Millennium v. Millennium Music, 33 F.Supp.2d 907, 49 U.S.P.Q.2d 1878, (D.Or., Jan 04, 1999) at for a summary of case law dealing with web sites and personal jurisdiction.

Q. Should the simple act of posting a web page (or registering a domain name) make one liable in any jurisdiction in which that page can be accessed? Consider the fact that anyone posting a page knows that it is available to a worldwide audience.

If you believe that such liability should ensue, how would you recommend handling the following case: a US company registers "," a site devoted to selling and trading music by a largely-forgotten 1980s pop group. A German company, LuftBalloons, has a valid German trademark in the name and considers the US registration infringing. Does your answer change if you know that the German site has a huge following in Germany? What if the US company was aware of this?

If you believe liability should only attach in a limited jurisdiction, how would you set the jurisdictional boundaries?

Foreign courts have exercised jurisdiction without regard to site of the host server or the domain registry. Some of them have applied an "effects approach" which involved consideration of the location in which the plaintiff had suffered harm and held that it had jurisdiction. The reasoning is that since it could establish that a local resident could access the Web site in the jurisdiction and was misled or deceived. While these cases have so far involved a resident defendant, they may have broader consequence and permit US entities to be sued abroad for harm caused by a "US" site in foreign location and also for an American company to sue abroad in order to protect its trademark. Burk, Jurisdiction in a World Without Borders, 1 Va. J.L. & Tech. 3 (Spring 1997).
Squatters and Pirates
The most frequent basis for litigation involves the practices of cybersquatting and cyberpiracy. These terms have not been carefully defined and yet almost every court around the globe has universally condemned both practices, whether under trademark law or some version of unfair trade practices. See Diane Cabell, Foreign Domain Name Disputes (1998).

"Cybersquatting" generally refers to the practice of stockpiling domain registrations in bulk for future resale to the general public. This activity can be innocent of any extortionist intent as with entities that auction or resell domain registrations as a service to a general public intimidated by the technical complexity of the TLD registrar interface. "Cyberpiracy" on the other hand, refers in particular to the registration of well-known names with the intent of deriving revenue from the sale of anotherís branded property. It also includes the act of registering names that are confusingly similar to a famous mark or well-known name for the purpose of generating traffic to a web site that would be ignored but for the userís confusion over the names. Examples of the latter include slight misspellings ( instead of Nike) or the use of different extensions ( is an adult sex site, is the page for the residence of the President). Drafting a consensus definition of cyberpiracy has been one of many hurdles in resolving domain disputes. See paragraphs 170 - 176, Final Report of the WIPO Internet Domain Name Process, WIPO Publication No. 92-805-0779-6 (April 30, 1999) at Courts have found both cybersquatting and cyberpiracy to be infringing practices.

A key issue has been whether registration of a domain alone (without any website activity) entails the necessary commercial use. Although many courts have claimed that mere registration alone does not constitute infringement, key early cases held that the registration in bulk coupled with an offer to sell provided the necessary commercial element to bring a successful claim of infringement against a cyberpirate. See Panavision International, L.P. v. Toeppen, 938 F. Supp. 616 (C.D. Cal. 1996), (denying motion to dismiss), partial summary judgment granted, 945 F. Supp. 1296 (1996) (against defendant Toeppen, partial summary judgment granted, 41 U.S.P.Q.2d (BNA) ¶ 1310 (Nov. 27, 1996) (for defendant Network Solutions, Inc.), summary judgment against defendant Toeppen affíd, 141 F.3d 1316, 46 U.S.P.Q.2d (BNA) ¶ 1511 (9th Cir. Apr. 17, 1998).

While a geographical address may be required to refer to a location, and thus may be "generic" and not likely to be registerable as a trademark owned by a business at that address, this assumes that the business takes its mark from a fixed address. With Internet addresses, the address is generally taken to match the businessí mark. Thus, "proximity" of the marks, i.e. the goods or services "labeled" with the mark, and how similar they are to each other is likely to be more important on the Internet than in the "real world." Dan L. Burk, Trademarks along the Infobahn: A First Look at the Emerging Law of Cybermarks, 1 RICH. J.L. & TECH., Paragraph 62 (Apr. 10, 1995).

The case of Avery Dennison Corp. v. Sumpton, 999 F. Supp. 1337, 46 U.S.P.Q.2d 1852 (1998) involved a cybersquatter. Sumpton claimed to be running a domain registration service and had registered 12,000 common surnames for resale to potential clients for use as e-mail addresses. Two of those registrations were and" These were found to be in violation of the Lanham Actís anti-dilution protection of famous marks (15 U.S.C. 1125©) despite the fact that Avery Dennison was the owner of the domain "" and therefore in no way prevented from marketing under its actual trademark (which is limited to the combination of both words). The 9th Circuit reversed that decision largely on the grounds that Avery did not qualify as a "famous" mark under the anti-dilution standards. Avery Dennison Corp. v. Sumpton, Case No. 98-55810, August 23, 1999.

Q. Surnames are ordinarily not registerable as marks unless they acquire secondary meaning. Is trademark practice on the Net reversing the basic principles of this law?

Q. Read the Avery Dennison case and the statute in the above paragraph. In particular, pay attention to the following Avery passage and its surrounding text:

The court holds that for purposes of the Act, a famous mark is "used in the ordinary course of trade" when (a) it is registered as a domain name by a registrant who is not otherwise identified by or associated with any of the commonly accepted meanings of the domain name, and (b) it is not used by the registrant as its own domain name, but rather is held by the registrant for sale or license to others. Under the reasoning of Avery, what justification would you offer for the ability of the domain name registrars, such as, to conduct their business?
Mere registration of a domain may not give rise to any trademark rights in a name. According to the Ninth Circuit U.S. Court of Appeals, "Registration with Network Solutions...does not in itself constitute 'use' for purposes of acquiring trademark priority." Brookfield Communications, Inc. v. West Coast Entertainment Corp., ----, 1999 WL 232014, 50 U.S.P.Q.2d 1545 [9th Cir.(Cal.) Apr 22, 1999] (NO. 98-56918)
Non-commercial Use  
Product Criticism is the URL for a site devoted to consumer complaints about Bally Total Fitness. District Judge Dean Pregerson reportedly refused to grant summary judgment to Bally on its claims that the site infringed on and diluted its registered mark. Additionally, the court reportedly invited the defendant to file his own summary judgment motion against Bally seeking dismissal of the claims against him. Bally Total Fitness Holding Corporation v. Andrew S. Faber, 29 F. Supp. 2d 1161 (C.D.Cal., Nov. 23, 1998). Other actions are pending in Arizona over a U-Hell website (U-Haul claiming trademark violation and libel) and in NY over (under anti-dilution claims). See generally Andrew Marlatt, Update/Companies Take Complaint Sites to Court, Internet World (November 16, 1998).
      Q. What about product information sites such as "" and "" where the domain holder is legitimately reselling a product or providing other support to product purchasers?

      Q. Those who post U.S.-based criticism web sites rely on the First Amendment to the U.S. Constitution to support their right to free speech. Do you think trademark owners have a right to insist that their trademark not be used in web site criticism? Should trademark law trump the First Amendment?

      Q. If you believe that the trademark owners have a right to insist that their name not be used in this fashion, how do you reconcile this with the fact that U.S. companies are allowed to use a competitorís trademarked name in advertising for purposes of product comparison?

        Free Speech
Political use of domain names has not fared so well. In two different cases, protestors used the names of well-known organizations as domain names leading to sites filled with material in moral and political opposition to the goals of the organizations. In Planned Parenthood Federation of America, Inc. v. Richard Bucci, d/b/a Catholic Radio, 42 U.S.P.Q.2d 1430 (S.D.N.Y. 1997), affíd. - F. 3d - (2d Cir., Feb. 9, 1998), the court held that the defendantís sales of books on the anti-abortion site was commercial use amounting to infringement. In Jews for Jesus v. Brodsky, 993 F. Supp. 282 (D.N.J., Mar. 6, 1998) there was even less commercial activity, so the court relied on the defendantís avowed purpose of intentionally misleading consumers who thought they were reaching the actual organization.
Q. If there were no commercial activity at all, what claim, if any, could the plaintiffs have brought against the domain holder for attempting to divert traffic

Reverse Hijackers
There is growing concern on the Internet about the abuse of non-commercial domain holders, particularly because they rarely have the financial resources to protect themselves in court.

Cases that have roused particular ire among Netizens involved children with domain registrations using their personal names. Although the sites themselves were clearly non-commercial (detailing such events as school baseball games and camp activities), impatient trademark counsel nevertheless sent cease and desist letters which gave rise to the accusation of reverse domain hijacking. The term refers to the abuse of trademark rights that occurs when a mark owner threatens a non-commercial domain holder with litigation thereby forcing the holder to undertake the financial burden of defending a fair use. In both the Pokey and Veronica cases, a grassroots effort of Net activists and resulting negative media coverage resulted in the mark owners reversing direction. See Beth Lipton, Archie Comics Fights Parent for Domain, CNet News, (Jan. 15, 1999).

Q. What is the real cost of defending a non-commercial domain registration against a litigous trademark owner?
Some legal theorists believe that the Internet will necessitate reconsideration of the very nature of intellectual property rights. As more and more words are grabbed for proprietary use, we are running out of words (as well as artistic expressions, methods and processes) and into more frequent collisions with our Constitution. See the collection of essays written by Lawrence Lessig, Berkman Professor of Entrepreneurial Legal Studies at Harvard Law School.

Commercial Use
Non-competing Use
Using traditional trademark principles, a New York district court dismissed claims for trademark infringement, false designation of origin, state infringement and unfair competition brought by toy retailer Toys ĎRí Us against a Massachusetts firearms dealer who operates out of his home and registered the domain name "gunsareus" for his Web site. Toys R Us, Inc. v. Feinberg, 26 F.Supp.2d 639 (S.D.N.Y., Oct 28, 1998). The Court dismissed the claims in part because Richard Feinbergís domain name does not use the toy retailerís trademarked "R." The Court further emphasized that guns are quite different from toys and that Feinberg uses his Web site to sell his wares primarily to other firearms dealers whom the Court considered to be sophisticated consumers unlikely to be confused by the name "gunsareus."

Interstellar Starship Services, Ltd. v. Epix, Inc., 983 F.Supp. 1331, 1997 WL 736486, 45 U.S.P.Q.2d 1304 (D.Or. Nov 20, 1997) (NO. 97-107-FR) involved a defendant (Starship) that had registered and used the domain name to publicize a theater group performing the Rocky Horror Picture Show. Plaintiff, Epix, Inc., a circuit board and computer program manufacturer owns the federally registered trademark Epix. While the trademark was found to be valid, the court determined that the use by defendant did not infringe on the plaintiffís trademark because the required element of likelihood of confusion was missing. The court emphasizes that the relevant likelihood of confusion "is confusion that affects the purchasing decisions of actual or prospective purchasers of the products of Epix, Inc." There is another unusual element to this case. The two firms say the word "epix" differently ("ep-ix" versus "e-pics"). One of the sites has an audio script that pronounces the word to further distinguish it from the other.

Note that often a likelihood of confusion analysis includes not only confusion that would lead to a mistaken purchase, but also confusion that would lead a potential consumer to believe that the product sold is affiliated with the trademark owner.

Competing Use
A series of Playboy cases, Playboy Enterprises, Inc. v. Calvin Designer Label, 985 F. Supp. 1218 (N.D. Cal. 1997) (temporary restraining order) injunction granted, 985 F. Supp. 1220 (1997); Playboy Enterprises, Inc. v. Asiafocus Intern., Inc. , 1998 WL 724000 (E.D.Va.,Apr 10, 1998) (NO. CIV.A. 97-734-A), have favored the trademark owner against use of the mark in the domain names (as well as in the website content and metatags) of competitors. These cases represent traditional trademark law applied in a traditional manner.

What if the competitor hasnít yet begun to use the domain? One of the cases that has found liability for use of a mark in a domain name even in the absence of any product or service was Green Products, Inc. v. Independence Corn By-Products Co., 992 F.Supp. 1070 (N.D. Iowa, Sept. 25, 1997). The court directed defendant to transfer ownership of a domain name to the plaintiff owner of an identical trademark. The court found that defendantís conduct was likely to infringe plaintiffís mark, and to cause consumer confusion, even though, as of the date of suit, defendant had not operated a website at "" The court rejected defendantís argument that its conduct would not confuse the public because defendantís site would prominently disclaim any relationship to plaintiff, and otherwise distinguish defendantís through comparative advertising. The court held that use of plaintiffís mark in a domain name would deceptively lure customers to defendantís site and thereby cause consumer confusion.

Q. What claim could a trademark owner bring against someone who registers the same word but simply sits on it to prevent the mark owner from accessing it?
One creative solution to conflict is an "intermediate home page" which involves the shared use of a domain name, either voluntarily or by stipulated judgment. Some are voluntary. See Others have been court ordered as where two companies were found to be jointly entitled to the domain name. Mattel v. Hasbro (Mattel, Inc. et al v. Hasbro, Inc. et al., Civil Action No. 96-7635, (U.S.D.C. C.D. Cal.). An introductory page at allows the user to link to either "" or ""

See the module Copyright in Cyberspace for a discussion of issues relating to the use of marks in the content of a web site such as links, frames and metatags.

Q. Is the Internet a single market place for purposes of determining confusion?

The fairly recent (1996) federal Anti-dilution Act, 15 U.S.C. 1125©, has special protections for famous marks and this is raising some of the most uncomfortable conflicts on the Net when it is used in an attempt to enforce US law beyond its territorial limits. The Act bans any new commercial use of a mark after it becomes famous, regardless of confusion, if the result is to diminish or tarnish the famous mark.

Where the famous mark is coined or particularly distinctive, the use of anti-dilution is rarely questioned. A typical example of an egregious situation is Toys "R" Us, Inc. v. Akkaoui , 1996 WL 772709, 40 U.S.P.Q.2d 1836 (N.D.Cal., Oct 29, 1996) (NO. C 96-3381 CW), where the domain name "" was used for a sexually oriented Internet shopping service. The court had no problem finding dilution by tarnishment of the famous mark.

Will the Anti-dilution Act be applied retroactively? An early North Carolina decision held that the domain owner of could keep it because it was obtained years before the plaintiff Gateway Computers became famous. Gateway 2000 v. Gateway.Com, Inc., 1997 U.S. Dist. LEXIS 2144, (E.D.N.C. 2/6/97). As yet unanswered is the application of the Act to domains that are registered after a mark became famous but before the Act took effect. Could the laundry detergent manufacturer displace someone who might have been selling nautical charts at since 1993?

Q. The actor, Leonardo DiCaprio, has applied for a trademark on his name, presumably because it has acquired secondary meaning as the source of certain entertainment services. Would his mark qualify as a famous one? If so, will he be able to prevent all other individuals with the same name from using it in any new enterprises? See the 9th Circuit's opinion in Avery.

Lower level domains and path designations

In Patmont Motor Works, Inc. v. Gateway Marine, Inc., 1997 WL 811770 (N.D.Cal.,Dec 18, 1997) (NO. C 96-2703 TEH) the use of plaintiffís trademark in the file path designation of a URL did not constitute trademark infringement. The reasoning is that file path designations (which are the characters to the right of the slash after the TLD) do not identify the origin of the web site but merely describes the siteís organization. For example, in the URL
Q. What about third and fourth level domains e.g.,; Should they be treated differently than higher level domains?

Other Name Rights
California is leading the way in protecting rights of publicity. Courts there have upheld those rights against domain holders who take advantage of the lure of celebrity names. In the Tim McGraw case, the court directed the defendant both (a) to stop operating websites at domain names he had registered with NSI which contained the names of various famous musicians and (b) to transfer those domain names back to the plaintiff musicians. Tim McGraw, et al. v. Jim Salmon, Case No. CV 98-2495R, 1998 U.S. Dist. Lexis 10987 (C.D. Cal., June 30, 1998). Legitimate fan sites, where no commercial activity takes place, are more problematic. In such situations, it might be in the celebrityís best interest to grant a suitable license. Presumably, rights of privacy might also be enforced in situations involving domains.

New litigation strategies
Trademark law principles force mark owners to aggressively enforce their rights within their territory, but that territory has now perhaps become as limitless as cyberspace itself. Not only must a mark owner like Disney keep track of any Internet use of that word, but also Disnee, Dysney, Walt Disney, Walter Disney, Walt Dinsnee, etc., etc., etc. And it must do so in each of the 260+ TLDs (,,, etc., etc., etc.). Since it is easy to do this online where simply entering the word as a search term will instantly locate most infringers, this places a significant burden on mark owners who must then seek redress in all these distant fora. While few people are shedding tears for famous brand owners with very deep pockets, often the issue isnít about money. A site that uses the domain name "Pocahontas" for pornography is a problem that cannot be ignored. Some new tactics may ease that load.

One of these new tactics is being used by Porsche Cars. It has brought an in rem action in Virginia against more than 80 NSI domain registrations, an action that unfortunately did not amuse the trial court. Porsche Cars North America, Inc. and Porsche AG, vs., et al., Civil Action No. 99-0006-A (E.D.Va, January 6, 1999). See complaint and order. Another maneuver is a Texas class action suit brought by hundreds of mark owners against a major cybersquatter. Omega Protein Corp. v. Flom, No. H-98-3114 (S.D.Tex. filed 9/18/98). The defendant had hoped to settle individually with each mark owner for something less than the cost of litigation, but this tactic will defeat that purpose.

See also the recent case involving Worldsport Networks Limited. Here the court held that it has the power to issue an order directing the non-party NSI to (a) transfer an infringing domain name (registered by the defendants) to the owner of the infringed mark, and to (b) refuse to register proscribed domain names by defendants in the future which will similarly infringe plaintiffís trademark. On NSIís application, the court modified its prior order so as to remove the orderís direction that NSI prevent "any party" from registering a name similar to plaintiffís trademark. This aspect of the order was found to impose an unreasonable burden on NSI and went beyond the permissible authority of the court to prevent defendants from engaging in infringing activities in the future. Worldsport Networks Limited v. Artinternet S.A. and Cedric Loison, Civ. Act. No. 99-CV-616, 1999 U.S. Dist. Lexis 6080; 1999 WL 269719 (E.D. Pa., April 28, 1999).

Technical Solutions
Technology may yet find a solution for much of the trademark conflict with domain names. A cyberpirate or infringer derives value for his domain by attracting web surfers who mistakenly believe they are accessing a site belonging to the true trademark owner. This happens most often when users enter the trademark term in a search engine or in the locator bar on the web browser. Search engines are now taking responsibility for directing the searcher to the legitimate trademark ownerís site, either by placing the legitimate site first among the search responses, by eliminating all fraudulent responses through filtering mechanisms or by establishing a search database that contains only the legitimate web sites. Centraalís Real Names technology, for example, is used by Look Smart, Inktomi, Netscape and Alta Vista. When a user types in a search term (e.g. White House), Real Names "resolves" that inquiry to official .gov web sites that have cleared a prior authentication review. Centraal Corporation homepage. In the quest for quicker, faster results, however, few users are aware that their searches are being constrained in any manner. Jonathan Zittrain, Keyword Obsolete, Wired Magazine (September 1998). In fact the difference between a true search engine (one that sends a robot out to digest and sort all web content in its path) versus a directory (where value judgements are made about content inclusion) may not even be well know to most Net users. Q. What claims might a commercial concern have concerning its placement on a search engine response results?

Q. Is the search engine substituting its judgment for that of a knowledgeable court?

Q. Should there be any duty of a filtered search engine to disclose its practices to the consumer?

Tags: Linking, Framing and Searching
  Links & Frames
The most distinctive feature of the World Wide Web is the ability to move instantly from one location to another. This is accomplished by embedding commands (tags) in the source code of a web site (known as "HREF" tags for Hypertext REFerence). When the user clicks on the link that appears on the source site, a command is sent to the webpage specified in the tag (the target site) and a request is made to send a copy of the material at that location back to the user and open it in the userís browser window. A text link is one which appears as text on the userís monitor. Text links are often underlined, but need not be. An image link is one that appears as a graphic. Tags can be programmed to bring all or simply a part of the other site to the userís browser. When the tags are designed to bring material from the target site and open it in conjunction with material from the source site, a practice that is often done when the target material is surrounded or framed by the source material, this can lead to litigation. Because the user remains at the source site, linking and framing may create the impression, at least for Net Newbies, that there is an authorized association between source and target. Commentators who have analyzed this issue have concluded that such links are most likely non-actionable under applicable copyright or trademark law. Martin J.Elgison and James M. Jordan III, Trademark Cases Arise from Meta-Tags, Frames. There are no judicial opinions in the US on text links as yet, but just as with domain names, trademark owners with deep pockets are quite willing to threaten litigation in order to control this practice. Carl Kaplan, Is Linking Always Legal? The Experts Aren't Sure, NY Times on the Web (August 6, 1999).

Copyright issues related to linking and framing are discussed in the module Copyright in Cyberspace. Trademark infringement and unfair trade practice grounds have usually been claimed as well although most such suits have settled before judgement.

The issue was originally raised by Ticketmaster when it objected to "deep links" from a Microsoft site that took surfers directly to the ticket order page thereby avoiding the expensive advertising on the Ticketmaster home page. Ticketmaster Corp. v. Microsoft Corp. (S.D. Cal. complaint filed Apr. 28, 1997), (settled). The fees being paid by the sponsors of the target site are based on the simple assumption that site visitors will see the ads as they navigate their way through the site. When source sites link to target sites in such a way that surfers are re-routed around the ads, that assumption is threatened. Not only does "deep linking" implicate advertising concerns, but it may also undermine the expectations of website authors. For example, deep linking to an article on a news site not only deprives the target site's sponsors of expected exposure, but it also inhibits the ability of the site's author to properly identify herself or otherwise to take credit for the article.

The related practice of "framing" may also serve to undermine web advertising. Ads are sometimes presented in "frames" that surround the web page itself. This allows advertisements or other framed information to remain constantly visible while the visitor navigates his or her way through the site. Controversy arises when both the source and target sites use frames. In such a case, the source site may use its own frames to obscure or cover up the target site's frames. Thus, a visitor following a link from a framed source site to a framed target site will see only the source site's frames, and hence ads. As with deep linking, the target site's sponsors are being deprived of the exposure for which they paid and trademark dilution may also occur. Washington Post Co. v. Total News, No. 97 Civ. 1190 (PKL) (S.D.N.Y.complaint filed Feb. 20, 1997) Complaint and Settlement.

Q. Courts recognize rights in the "look and feel" of a product to the extent those elements are not functional. This is the concept of "trade dress." See Two Pesos, Inc. v. Taco Cabana, Inc., 505 U.S. 763 (1992) Opinion. Is framing a possible violation of trade dress?

There are some technical solutions to unwanted linking and framing, which may be the reason why so little litigation has developed in this area. Target sites can codes their pages to open without being framed, deep pages can be programmed to accept traffic only from the home page, and advertising can always be moved to the pages that catch the most eyeballs. Ticketmaster was able to redirect users deep-linking from MSís Sidewalk long before their legal battle was over. Technical solutions may also prove much less expensive and time-consuming than litigation, as well.

Meta tags & Searches
Locating information on the Internet is still largely a text-based exercise, although image and sound directories are expanding rapidly. Many users guess at the location of a vendor, usually by typing www.vendorname in the browserís Uniform Resource Locator bar. This is one reason why getting a domain to match oneís trademark is so important and why owning a domain name that is a short, generic word, such as "" is the digital equivalent of a goldmine.

Another method to find data is to use one of the many search engines. Some search engines are actually just directories containing a limited set of pre-selected URLs. Others utilize robots that continually traverse the Net, reading and indexing every web page in their paths. To speed up search technology, many Ďbots are directed to digest merely the pageís meta tags which are simply a selection of keywords that indicate the type of material that is found on the page. The keywords used in meta tags are selected by the web page owner. Meta tags are not visible to the user, however.

One area of litigation involves using anotherís trademark in the meta tags of a competitorís web site. In the "moviebuff" case, Brookfield, an appellate court held that such use led to "initial interest confusion" which occurs when a person seeking a trademark product is deterred from finding it due to distraction during a search by the appearance of anotherís product. This results in an actionable loss of goodwill.

Playboy Enterprises was able to enjoin use of its trademark terms "playboy" and "playmate" in the meta tags of a competing adult magazine [Playboy v. Calvin, below], but was not able to bar such use by a former playmate where the terms were used as legitimate descriptors of her status. Playboy Enterprises v. Welles, 47 U.S.P.Q.2d 1186, 7 F.Supp.2d 1098 (1998), affirmed, on appeal to the 9th Cir. No.98-55911, (October 20, 1998) Playboy then added a new claim to its lawsuit: trademark counterfeiting. Finally, Welles countersued for defamation and unfair-business practices.

Playboy also lost the first round in its complaint against the owners of the Excite search engine for selling banner ads on the page where the results of searches for the words "playboy" and "playmate" were posted. The court distinguished this situation from the earlier cases involving meta tags on the basis that the terms were used in their generic sense (playboy), not in their trademark format (Playboy®). Brian McWilliams, Excite Wins Playboy Lawsuit, (July 15, 1999). Playboy Enterprises, Inc. v. Netscape Communications --- F.Supp.2d ----, 1999 WL 428233 (Jun 24, 1999).

Q. One can analogize the practice of using meta tag keywords to place higher on search engine hit lists to the practice of starting oneís company name with as many "A"s as possible to ensure top billing in a Yellow Pages listing. When you open the Yellow Pages to find the number for Worldwide Airlines, you may decide instead to call AAAAmerican Air, since you see them first and donít take the time to page through to Worldwideís number.

Q. What justifications can you come up with for allowing the meta tags practice?

Q. What reasons would you use for invalidating it?