"When I use a word," Humpty Dumpty said in rather a scornful tone, "it means just what I choose it to mean — neither more nor less."

"The question is," said Alice, "whether you can make words mean so many different things."

"The question is," said Humpty Dumpty, "which is to be master — that's all."

Lewis Carroll, Alice’s Adventures In Wonderland and Through The Looking Glass 163 (Gray ed. 1971).

Business people taking their names and brands into the world of online commerce, and the lawyers who advise them, can empathize with Alice. In opening an internet browser, it is easy to feel that one has stepped through the looking glass: the rules have all changed, words all seem to mean something different, and the battle to determine who is master of those words in the online world remains very much in doubt.

But there is good news! Rules governing the use and protection of trademarks in the online world do exist, and they are gradually becoming more clear and settled. Even better, while the new context has introduced some new wrinkles, by and large the underlying principles applied are the traditional and time-tested concepts of trademark law which evolved way back in the twentieth century, on the other side of the looking glass.

This paper will begin with a brief review of those fundamental trademark principles, followed by an analysis of how they are applied to issues which uniquely arise in the online context, including domain names, metatags and keywords, linking and framing, and related issues.

1. Trademark Basics

   This section presents an overview of basic trademark principles, which is necessarily summary and incomplete in nature. For a thorough treatment, consult Professor McCarthy’s authoritative work, McCarthy On Trademarks and Unfair Competition (4^th ed. 2000), which is available on CD-ROM as well as on dead-tree media.

   1. Definition of Trademark
      
      
      1. A trademark is "any word, name, symbol, or device, or any combination thereof, used by a person . . . to identify and distinguish his or her goods, including a unique product, from those manufactured or sold by others and to indicate the source of the goods, even if that source is unknown." Lanham Act § 45, 15 U.S.C. 1127.
         
         
      2. Service marks are of exactly the same character, except that they are marks applied to services rather than goods. Id. In this manuscript, the terms "trademark" or "mark" are used interchangeably to refer to both trademarks and service marks.
         
         
      3. In theory, any symbol can function as a trademark, including not only words, but graphic designs and logos, overall label and package appearance (known as "trade dress"), product shape, and even such non-traditional symbols as sounds and aromas.
         
         
   2. Establishing Trademark Rights
      
      
      1. In the United States, trademark rights arise from use, that is, from the actual application of the mark to goods or services in commerce.
         
         
      2. Registration of the mark with the U.S. Patent and Trademark Office provides many benefits, such as presumptions of ownership and nationwide use as of the date claimed in the application, Lanham Act 33, 15 U.S.C. 1115, and constructive notice to others of the registrant’s claim of ownership, Lanham Act § 22, 15 U.S.C. § 1072. An application for registration may be filed prior to actual use based on a bona fide intent to use the mark, Lanham Act § 1(b), 15 U.S.C. § 1051(b), but no registration will issue until the mark is actually used. Lanham Act 1(d), 15 U.S.C. 1051(d).
         
         
      3. As a result, despite the prevalence of registration of marks in the U.S., trademark rights can still be obtained without registration simply based on appropriate use of the symbol in commerce as a trademark.
         
         
   3. What Is Trademark Use?
      
      
      1. The use of a word or other symbol sufficient to create trademark rights and support registration must be a "trademark use," that is, the symbol must be used as a source identifier for the goods or services to which it is applied, and the relevant consuming public must recognize it as such, based on either inherent or acquired distinctiveness of the mark.
         
         
      2. "Inherent distinctiveness" means that the symbol is of such a nature that the law will presume that consumers recognize it as a source identifier, even if they have never seen it before. "Acquired distinctiveness" means that, though not inherently distinctive, the mark has been advertised and used widely enough that consumers have come to recognize it as a source identifier anyway.
         
         
      3. Deciding whether a particular mark is or is not inherently distinctive can be a tricky business. To facilitate the analysis, marks are "classified in categories of generally increasing distinctiveness; . . . they may be (1) generic; (2) descriptive; (3) suggestive; (4) arbitrary; or (5) fanciful." Two Pesos, Inc. v. Taco Cabana, Inc., 505 U.S. 763, 768 120 L. Ed. 2d 615, 112 S. Ct. 2753, 23 U.S.P.Q.2d 1081 (1992).
         
         
   4. The Spectrum Of Distinctiveness
      
      
      1. Arbitrary or fanciful marks are those which by their nature present no plausible basis for consumers to think that they are being used for any non-trademark purpose, such as merely to describe some characteristic of the goods. Since such terms have no apparent purpose other than to symbolize the source of the goods to consumers, it is presumed that they do exactly that. Fanciful marks have this quality because they have no independent meaning whatsoever. Examples are KODAK for film and EXXON for gasoline. Arbitrary marks have this quality because, even though they have a pre-existing recognized meaning, that meaning has no logical connection whatsoever to the goods to which they are applied, such as APPLE for computers or AMAZON for bookselling services.
         
         
      2. Suggestive marks have some relation to the goods to which they are applied, but fall short of directly describing a quality or characteristic of the goods, and instead require some imagination on the part of the consumer to make a connection. Examples are TURTLE WAX for car polish (indirectly suggesting that the product produces a finish on cars as hard as a turtle’s shell) or MUSTANG for automobiles (indirectly suggesting that the car has the power and speed of a mustang).
         
         
      3. Descriptive marks directly describe an ingredient, quality, characteristic, function, feature, purpose or use of the goods. Unless consumers are already familiar with the word as a brand name for this product, they might just as likely assume that the term was simply being used to describe the product, not to identify its source. Examples are SHARP for cutlery or SOFT for facial tissues.
         
         
      4. Generic terms are those which serve as a common name for a type, category or sub-category of goods, such as "furniture" and "sofa." The category invoked by a generic word or phrase can be very narrow, so long as it is clear that the principal meaning of the term in the minds of the consuming public is to signify a type of goods, and not goods coming from a particular producer or seller. For example, DIET CHOCOLATE FUDGE SODA has been held to be generic for, well, diet chocolate fudge soda. A.J. Canfield Co. v. Honickman, 808 F.2d 291 (3^rd Cir. 1986).
         
         
   5. Consequences For Registrability And Strength
      
      
      1. The reality is that the above four (or five, if you count "arbitrary" and "fanciful" as separate) categories are actually a somewhat arbitrary structure imposed on what is really a graduated continuum of more or less distinctive words and symbols. Nonetheless, for purposes of the law, a proposed mark will usually be assigned to one or another of the categories, and the category to which it is assigned has significant consequences for the registrability and enforceability of the mark. Note that where a particular term falls in the spectrum cannot be determine except in the context of the goods to which that term is applied. APPLE is generic and entitled to no protection as applied to fruit, but it is arbitrary and entitled to the highest protection as applied to computers.
         
         
      2. Fanciful, arbitrary and suggestive marks are often referred to as "inherently distinctive", and are registrable as trademarks simply upon proof that they have been or are intended to be used on specific goods.
         
         
      3. Descriptive terms, by contrast, do not give rise to trademark rights and may not be registered unless the owner provides satisfactory proof that, notwithstanding the original descriptive meaning of the word, the owner has used, advertised and promoted the word as a mark for its goods to such an extent that a substantial portion of consumers have come to recognize it as a symbol of source anyway. This is often called "acquired distinctiveness," or, in somewhat more arcane terminology, "secondary meaning" (the acquired trademark meaning of the term being "secondary" to the primary descriptive meaning).
         
         
      4. A term that is generic for the goods to which it is applied is not registrable under any circumstances. Even if a term has already been registered, even for a significant period of time, the registration is still vulnerable to cancellation if the challenger can establish that the term is or has become generic for the goods to which it is applied.
         
      5. "Non-traditional" trademarks sometimes receive specialized treatment. For example, a single color standing alone (as opposed to colors in a specific combination or configuration) are never inherently distinctive, and can function as a trademark only upon proof of acquired distinctiveness. Qualitex Co. v. Jacobson Prods. Co., 514 U.S. 159 (1995). Similarly, the shape or appearance of a product itself (as opposed to a label or package) must be proved to carry secondary meaning before it can receive trademark protection. Wal-Mart Stores, Inc. v. Samara Bros., Inc., ____ U.S. ____, 120 S.Ct. 1339, 146 L.Ed.2d 182 (2000).
         
         
   6. Proving Infringement
      
      
      1. Trademark rights do not necessarily confer exclusive ownership of a mark. Rather, what is owned is the right to prevent others from using the same or a similar mark in such a way as to create a likelihood of confusion among consumers as to the source or sponsorship of the goods or services involved. Once the existence of trademark rights has been established, the owner of a senior mark may halt a junior use only if use of a junior mark creates a likelihood of confusion among consumers.
         
         
      2. Likelihood of confusion means that consumers seeing the junior mark are likely to believe incorrectly that the goods on which the junior mark appears come from, or are authorized or approved by, the owner of the senior mark, or that there is some connection between the two. Once priority is established, this question of likelihood of confusion is the dispositive issue in virtually every infringement case.
         
         
      3. The courts employ a multi-factor test to decide whether there is a likelihood of confusion in a particular case. Such a test was first articulated in 1938 in § 731 of the Restatement of Torts. The first judicial incarnation was in Polaroid Corp. v. Polarad Electronics Corp., 287 F.2d 492, (2d Cir. 1961), cert. denied, 368 U.S. 820 (1961). Now known as the Polaroid test, it has spawned cousins in every other federal circuit. Formulations vary somewhat from court to court, but generally focus on factors such as the following:
         
         
         1. the strength of the senior mark;
            
            
         2. the degree of similarity between the two marks;
            
            
         3. the proximity of the competing products in the marketplace;
            
            
         4. the likelihood that the senior owner will "bridge the gap" between unrelated goods;
            
            
         5. the existence of any actual confusion;
            
            
         6. the junior user’s intent in adopting its mark;
            
            
         7. the quality of the junior user’s product; and
            
            
         8. the sophistication of the buyers of the respective goods.
            
            
      4. The courts generally acknowledge that no single factor is determinative, not all factors need be present for confusion to be likely, and all factors must be considered in "flexible interplay." Nonetheless, the first factor in every version of the likelihood of confusion test is the strength of the senior mark. Most courts recognize this as one of the most important factors, and some have explicitly held that proof of a very strong and well-known senior mark, coupled with evidence of actual confusion of consumers by use of the junior mark, is sufficient to make out a per se case of infringement without consideration of any of the other factors at all. See, e.g. Sara Lee Corp. v. Kayser-Roth Corp., 81 F.3d 455 (4^th Cir. 1996).
         
         
   7. Dilution
      
      
      1. In the early part of the twentieth century, there was more and more discussion of the idea that the system of trademark protection described above, dependent as it was on the presence or absence of likelihood of confusion between similar marks, did not fully protect all the legitimate interests of trademark holders. See, e.g., Schechter, "The Rational Basis of Trademark Protection," 40 Harv. L. Rev. 813 (1927), reprinted in 60 Trademark Rep. 334 (1970).
         
         
      2. The result was the evolution of the concept of dilution of trademarks, intended to provide a remedy when use of a competing mark caused injury to a senior mark even though there was no likelihood of confusion of consumers, because the goods were unrelated or for other reasons. Beginning in the 1940s, a majority of states enacted anti-dilution statutes. Federal legislation did not come until 1996, with the Federal Trademark Dilution Act, which added a new § 43(c) to the Lanham Act, 15 U.S.C. § 1125(c).
         
         
      3. 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." 15 U.S.C. § 1127. There are thought to be two ways such dilution can occur: blurring and tarnishment.
         
         
         1. "Blurring" is based on the idea that the use of a famous brand on unrelated goods can damage the value of the brand regardless of whether there is any customer confusion in a trademark sense. For example, consumers who see BUICK toothpaste might be very unlikely to think it was made or approved by the auto maker, but the distinctiveness of the BUICK automobile brand could be undermined anyway. Whenever consumers were exposed to the BUICK name, they would have to stop and think whether it was the car maker or the toothpaste maker that was using it. Additional third-party uses of BUICK on other products could eventually have the result of substantially undercutting the source-identifying power of the BUICK mark for automobiles.
            
            
         2. "Tarnishment" results from the use of a famous mark on a product with unsavory or inappropriate connotations given the senior use. The most frequent example is the use of a well-known mark in connection with sexually-explicit products or services, but any use which tends to undercut the esteem in which the mark is held by the consuming public could give rise to dilution by tarnishment.
            
            
      4. The federal law only worked a "quasi-pre-emption" of the existing state dilution statues. State dilution laws retain validity except where the allegedly diluting mark is federally registered, in which case federal dilution law is the sole remedy. 15 U.S.C. § 1125(c)(3). While the basic dilution concept is the same under federal and most state laws, some state statutes did not require that a mark be famous as a condition of dilution protection, an important limitation under the federal law.
         
         
      5. Dilution represents a substantial departure from traditional trademark doctrine, which historically has focused on protecting not so much the rights of the trademark owner, but the right of the public not to be confused or mislead, and therefore the right of the trademark owner to a non-confused public. As the First Circuit has observed, with dilution "an entirely different issue is at stake–not interference with the source signaling function but rather protection from an appropriation of or free riding on the investment [the mark owner] has made in its [mark]." I.P. Lund Trading ApS v. Kohler Co., 163 F.3d 27 (1st Cir. 1998).
         
         
2. The Online World And The Importance Of The Commercial Context
   
   As the above discussion makes clear, under traditional trademark doctrine the commercial context of any particular trademark use is critical. The nature of the goods and services to which a mark is applied, the channels of commerce in which those goods and services are advertised and sold, the geographic region in which advertising and sale takes place, and the identity of and characteristics of the buyers who are targeted, all function to demarcate the limits of the rights acquired by the trademark owner. Material differences between senior and junior marks in any of these dimensions may arguably be sufficient to preclude a likelihood of confusion.
   
   Dilution, at least for the owners of famous marks, expands the rights flowing from the trademark in some ways, conferring the ability to stop non-confusing competing uses which nonetheless blur or tarnish the public’s perception of the famous mark. Even with dilution claims, however, commercial context provides important limits. Fame of the senior mark may be limited to a particular geographic region, e.g., Wawa Dairy Farms v. Haaf, 40 U.S.P.Q.2d 1629 (E.D. Pa. 1996), aff’d, 116 F.3d 471 (3d Cir. 1997) (WAWA held to be famous in five-state region), or to a specialized group of consumers, e.g., Mead Data Cent., Inc. v. Toyota Motor Sales, Inc., 875 F.2d 1026 (2d Cir. 1989) ("[S]uch distinctiveness as LEXIS possesses is limited to the narrow field of attorneys and accountants.").
   
   When marks are used on the internet, the commercial context can be very different, and sometimes virtually non-existent. An inherent characteristic of the internet is its ability to facilitate access to information by disparate groups of people and businesses. It provides such access without respect to geography, and the information provided often has a much less defined context than in other media.
   
   Even the way in which information is accessed online often strips that information of context. Before the internet and the powerful search engines developed for it, a potential buyer of goods or services almost always conducted a content-based and context-rich search for information: stores, catalogs, product brochures, even media advertisements all provided plenty of signals as to at least the general nature of the products and services being discussed. The internet, by contrast, makes possible a literal search for keywords, and returns results stripped of context. Often, the exact nature of a site on the web is not apparent at all until a user makes the decision to click on a URL and actually visit a site.
   
   The result is that users of the internet are often bombarded with words, names and brands which lack much of the context on which traditional trademark law depends. That lack of context means that many traditional anchor points for trademark analysis (relatedness of goods and services, channels of trade, and so on) can be underdeveloped or entirely absent in the on-line context. As the following discussion of various specific internet-related issues illustrates, courts and legislatures have struggled in various ways to find a reasonable middle ground between the two extremes of finding that everything is confusing, or that nothing is.
   
   
3. Domain Names
   
   Domain names present perhaps the clearest example of how the internet can eliminate context and thereby increase the potential for a clash of trademark interests which would likely never have occurred in more traditional channels of commerce. When a web user sees a reference to the domain name www.clue.com, does it refer to a well-known board game, or to a Colorado systems integrator called Clue Computing, Inc.? While no individual customer of either party would be likely to come into contact with the name in the non-silicon world except in a context where it was very clear which of the two entities was involved, a list of domain names on the internet gives no clue (pardon the pun).
   
   To compound the problem, self-help is not an effective remedy. Cybersquatting is virtually impossible to control by pre-emptively registering domain names that might be of interest, because the possible variations are limitless. See, e.g., PaineWebber, Inc. v. wwwpainewebber.com, No. 99-0456-A (E.D. Va. 1999) (entering injunctive relief against cybersquatter who registered "wwwpainewebber.com," with no period after the www).
   
   As a result, a considerable body of case law has already evolved in the area, as well as specialized federal legislation and an entire system of private dispute resolution.
   
   
   1. Is A Domain Name A Trademark?
      
      
      1. The mere reservation of a domain name does not by itself constitute trademark use. See, e.g., Brookfield Communications, Inc. v. West Coast Entertainment Corp., 174 F.3d 1036 (9th Cir.1999) ("mere registration of a domain name [is] not sufficient to constitute commercial use for purposes of the Lanham Act"); Juno Online Services v. Juno Lighting, Inc., 979 F. Supp. 684 (N.D. Ill. 1997) ("The mere ‘warehousing’ of the domain name is not enough to find that defendant placed the mark on goods or ‘used or displayed [the mark] in the sale or advertising of services’ as required.")
         
         
      2. It is nonetheless clear that a domain name can be used as a trademark, and when it is so used, trademark law applies. See generally Teletech Customer Care Management v. Tele-Tech Co., Inc., 977 F. Supp. 1407 (C.D. Cal.1997); Green Products Co. v. Independence Corn By-Prods. Co., 992 F. Supp. 1070 (N.D. Iowa 1997); Playboy Enterprises, Inc. v. Calvin Designer Label, 985 F. Supp. 1220 (N.D.Cal.1997); Cardservice Int'l v. McGee, 950 F. Supp. 737 (E.D. Va.), aff'd, 129 F.3d 1258 (4th Cir.1997). Some courts have even suggested that simply using the domain name is enough. Juno Online Services v. Juno Lighting, Inc., 979 F. Supp. 684 (N.D. Ill. 1997) ("However, the ‘use in commerce’ requirement would only be fulfilled if defendant were to use the Internet."
         
         
      3. Things become a bit more murky when one attempts to discern exactly where the line is between use in commerce of a domain name sufficient to satisfy the requirements of the Lanham Act, and non-commercial use (or non-use). In Panavision Int’l, L.P. v. Toeppen, 141 F.3d 1316 (9^th Cir. 1998), the defendant argued that he couldn’t be liable for infringement or dilution because he had done nothing but reserve the domain name in dispute, and therefore had engaged in no commercial use of it. The court found, however, that that the defendant, who had registered a multitude of names embodying famous marks, was engaged in the business of arbitraging domain names, and therefore had made commercial use of the domain name in the course of that business. See also Intermatic Inc. v. Toeppen, 947 F. Supp. 1227 (N.D. Ill. 1996.) ("Toeppen’s intention to arbitrage the ‘intermatic.com’ domain name constitutes a commercial use. At oral argument Toeppen’s counsel candidly conceded that one of Toeppen’s intended uses for registering the Intermatic mark was to eventually sell it back to Intermatic or to some other party. Toeppen's desire to resell the domain name is sufficient to meet the ‘commercial use’ requirement of the Lanham Act.")
         
         
      4. By contrast, the court in HQM, Ltd. v. Hatfield, 71 F.Supp.2d 500 (D.C. Md. 1999), refused to find commercial use simply based on registration of a "dot com" name: "[T]he .com designation does not by itself constitute commercial use. To hold otherwise would create an immediate and indefinite monopoly to all famous marks holders on the Internet, by which they could lay claim to all .com domain names which are arguably ‘the same’ as their mark. The Court may not create such property rights-in-gross . . .").
         
         
      5. At least one court has held that prior use of a trademark does not support use of that mark with ".com" added. In Data Concepts, Inc. v. Digital Consulting, Inc., 150 F.3d 620 (C.A.6 1998), plaintiff allegedly began using "dci" as a trademark in 1982. Defendant obtained a federal registration of DCI in 1987. Plaintiff then obtained and began using the domain name "dci.com" in 1993. Citing the traditional "tacking" principle that "the use of an earlier mark can be tacked onto the use of a subsequent mark only if the previously used mark is ‘the legal equivalent of the mark in question or indistinguishable therefrom’ such that consumers ‘consider both as the same mark,’" Van Dyne-Crotty, Inc. v. Wear-Guard Corp., 926 F.2d 1156, 1159 (Fed. Cir. 1991), the Sixth Circuit held that plaintiff could not tack its use of dci.com to its trademark use of "dci" because the two were not "indistinguishable." As a result, plaintiff was deemed to have commenced its use of "dci.com" in 1993, when it obtained the domain name, and defendant was deemed to be the senior user by virtue of its 1987 trademark registration.
         
         
   2. Likelihood Of Confusion From Use Of Domain Names
      
      
      1. Here is where the internet rubber meets the road. Courts have wrestled with the enhanced potential for confusion presented by the context-thin internet environment, with varying and, to some extent, inconsistent results.
         
         
      2. Some courts have discounted any confusion arising merely from the fact that both parties offer services on the internet. See Bally Total Fitness Holding Corp. v. Faber, 29 F.Supp.2d 1161 (C.D. Cal. 1998) ("The fact that the parties both advertise their respective services on the Internet may be a factor tending to show confusion, but it does not make the goods related. The Internet is a communications medium. It is not itself a product or a service.") Courts of this view tend to find no likelihood of confusion based simply on the momentary "confusion" experienced by a user upon seeing a domain name similar to the mark of another, especially if that confusion is immediately dispelled when the user reaches the site and sees clearly that it offers unrelated products or services. TeleTech Customer Care Management (California), Inc. v. Tele-tech Co., Inc., 977 F. Supp. 1407 (C.D. Cal. 1997) ("TeleTech thus far has only demonstrated an initial confusion on the part of web browsers using the domain name "teletech.com" but finding the Defendant's website. This brief confusion is not cognizable under the trademark laws.")
         
         
      3. Other courts have seemed to suggest exactly the opposite, however, that presence on the internet by itself supports a likelihood of confusion. See, e.g., Planned Parenthood Federation Of America, Inc. v. Bucci, 42 U.S.P.Q.2d 1430 (S.D.N.Y. 1997), aff’d mem. 152 F.3d 920 (2d Cir.), cert. denied, 119 S.Ct. 90 (1998) ("Because plaintiff’s web site and defendant’s web site are both on the Internet, the parties are vying for users in the same ‘market.’"); Intermatic Inc. v. Toeppen, 947 F. Supp. 1227 (N.D. Ill. 1996) ("Both parties are attempting to establish a presence on the Internet through the creation of a web page. The distribution channel in this case is cyberspace. As consumers ‘surf the net’ they seek out information on a plethora of subjects or companies. Companies around the globe are scrambling to establish their presence on the Internet. It is axiomatic that companies seek to register their trademarks as domain names so that consumers can easily find information about them or their products and services.")
         
         
      4. Some courts have relied on the doctrine of "initial interest confusion," see, e.g., Mobil Oil Corp. v. Pegasus Petroleum Corp., 818 F.2d 254 (2d Cir.1987) (trademark similarity which causes a potential purchaser to be exposed to or consider the defendant’s goods or services is sufficient to support a trademark claim, even if the consumer is disabused of any confusion prior to actually making a purchase decision), to find trademark liability in the use of confusingly similar domain names. Interstellar Starship Services, Ltd. v. Epix Inc., 184 F.3d 1107 (9^th Cir. 1999) (despite differences in goods and services offered, a user expecting to find a website sponsored by plaintiff might "decide to give [defendant]’s services a try, permitting [defendant] to capitalize on the goodwill [plaintiff] developed in its trademark--even if the customer is never confused about [plaintiff]’s lack of connection to" the web site).
         
         
   3. Dilution And Domain Names
      
      
      1. Because of the conceptual difficulties in establishing likelihood of confusion from domain name use in some contexts, the dilution laws have provided a handy, though not universally successful, tool for attacking domain names embodying a famous mark. Its utility for this purpose was in fact anticipated by the authors of the federal dilution legislation. See 141 Cong. Rec. § 19312-01 (daily ed. Dec. 29, 1995) (statement of Sen. Leahy) ("[I]t is my hope that this anti-dilution statute can help stem the use of deceptive Internet addresses taken by those who are choosing marks that are associated with the products and reputations of others.")
         
         
      2. For examples of the use of dilution to halt the use of domain names embodying famous marks in the absence of confusion, see Intermatic Inc. v. Toeppen, 947 F. Supp. 1227 (N.D. Ill. 1996); Panavision Int’l, L.P. v. Toeppen, 141 F.3d 1316 (9^th Cir. 1998). In Panavision, the Ninth Circuit held that, even though there might be neither blurring nor tarnishment, defendant’s reservation and attempted sale of a domain name embodying plaintiff’s famous mark nonetheless diluted it, saying, "To find dilution, a court need not rely on the traditional definitions such as ‘blurring’ and ‘tarnishment.’. . . [P]otential customers of Panavision will be discouraged if they cannot find its web page by typing in ‘Panavision.com,’ but instead are forced to wade through hundreds of web sites." This effect, says the court, "lessens the capacity of Intermatic to identify and distinguish its goods and services by means of the Internet," and therefore is diluting.
         
         
      3. Dilution does not eliminate the problem of showing trademark use by the defendant, however. The requirement that the diluting use be a "use in commerce" has been held to "require[] the defendant to be using the trademark as a trademark, capitalizing on its trademark status." Avery Dennison Corp. v. Sumpton, 189 F.3d 868 (9^th Cir. 1999). As a result, the commercial use issues discussed above remain.
         
         
      4. The easiest road is available to those who can show tarnishment of a famous mark. Few courts would have trouble concluding that a porn site’s use of the domain "candyland.com" tarnishes the well-known children’s board game. Hasbro, Inc. v. Internet Entertainment Group, Ltd., 1996 WL 84853 (W.D. Wash. 1996).
         
         
   4. Anticybersquatting Consumer Protection Act
      
      
      1. In late 1999, section 43 of the Lanham Act, 15 U.S.C. § 1125, was amended to add a new paragraph (d) which explicitly defined and provided a remedy for "cyberpiracy."
         
         
      2. A person is liable for cyberpiracy if, without regard to the goods or services of the parties, that person
         
         
         1. registers, traffics in or uses a domain name that is identical or confusingly similar to a mark that was distinctive at the time the domain name was registered, and
            
            
         2. has a bad faith intent to profit from that mark, including a personal name which is protected as a mark under this section
      3. If the mark was famous when the domain name was registered, then the domain name need not necessarily be identical or confusingly similar to the mark, but need only be "dilutive" of it.
         
         
      4. The statute provides a list of nonexclusive factors which a court may consider in determining bad faith, such as whether the person has his own trademark rights in the name, and also provides that bad faith may not be found where the person had "reasonable grounds to believe that the use of the domain name was a fair use or otherwise lawful."
         
         
      5. Remedies include
         
         
         1. forfeiture of the domain name or transfer to the plaintiff, 15 U.S.C. § 1125(d)(1)(C),
            
            
         2. damages, defendant’s profits and costs, 15 U.S.C. 1117(a), and
            
            
         3. as an alternative to damages and profits, statutory damages "in the amount of not less than $1,000 and not more than $100,000 per domain name, as the court considers just." 15 U.S.C. 1117(d)
            
            
      6. The law also provides for in rem jurisdiction over the domain name itself if the owner cannot be found with the exercise of due diligence. 15 U.S.C. § 1125(d)(2)(A). In an in rem action, remedies are limited to cancellation or transfer of the offending mark. 15 U.S.C. § 1125(d)(2)(D)(i). Though it is difficult to reconcile with the purpose of an in rem action, and is not compelled by the language of the statute, the courts have held that an in rem action may only be maintained if there is proof of bad faith by the registrant. Harrods Ltd. v. Sixty Internet Domain Names, 2000WL 1175103 (E.D. Va. 2000); BroadBridge Media v. Hypercd.com, 2000 WL 959715 (S.D.N.Y.2000).
         
         
      7. In addition to trademarks, the Act also created liability for registering a domain name consisting of the name of another living person, or a name substantially and confusingly similar thereto, without that person’s consent, with the specific intent to profit from such name by selling the domain name for financial gain to that person or any third party. 15 U.S.C. § 1129(1)(A).
         
         
   5. ICANN Uniform Domain Name Dispute Resolution Policy
      
      
      1. As an alternative to litigation, the Internet Corporation for Assigned Numbers and Names (ICANN) has adopted a Domain Name Dispute Resolution Policy. Details are available at http://www.icann.org/udrp/udrp.htm.
         
         
      2. Any person who claims that a domain name registered by another is identical or confusingly similar to a mark owned by that person may initiate an administrative dispute resolution proceeding, which the registrant is required to consent to at the time of the registration or renewal of the domain name. Three elements must be shown to obtain relief:
         
         
         1. the domain name is identical or confusingly similar to a trademark or service mark in which the complainant has rights; and
            
            
         2. the registrant has no rights or legitimate interests in respect of the domain name; and
            
            
         3. the domain name has been registered and is being used in bad faith
            .
      3. The dispute is arbitrated before a third party arbitration provider approved by ICANN. Four organizations have presently received ICANN approval, the most prominent and active of which is the World Intellectual Property Organization (WIPO).
         
         
      4. The procedure applies to all names in the .com, .net and .org top-level domains. In addition, registrars for country-level domains can adopt the procedure, and some have done so (.nu, .tv, and .ws, for example).
         
         
      5. Relief under the dispute resolution procedure is limited to cancellation or transfer of the domain name. Moreover, any party can choose to commence a judicial action at any time before or after the proceedings are complete, and in that case the registrar will not cancel or transfer the domain name except on order of the court.
         
         
      6. Because there is no formal appeal process, and because the individuals serving as arbitrators may be of widely varying background and experience, arbitration decisions reported to this point have been inconsistent, to say the least, and sometimes they conform only loosely, if at all, to applicable legal principles.
         
         1. One arbitrator recently ordered the registrants of the "Barcelona.com" domain, who had used it, most would say legitimately and in good faith, for their city directory web service, to transfer the domain to the City of Barcelona, Spain, on the dubious premise that visitors to the domain would expect it to be an official city site. See arbiter.wipo.int/domains/decisions/html/d2000-0505.html (All WIPO arbitration decisions are available at the WIPO Arbitration and Mediation Center website, arbiter.wipo.int.)
            
            
         2. Another arbitrator reportedly concluded that a domain name was confusingly similar to the challenger’s registered mark, that the registrant had no rights or legitimate interests in the name, and that the registrant obtained the name in bad faith, but nonetheless denied relief because there was no evidence that the registrant had used the name in bad faith, primarily because the registrant had failed to respond to the administrative challenge at all. A media account of the decision may be found at www.globetechnology.com/archive/gam/News/
            20000621/RLOBL.html
            
            
      7. The primary advantage of the procedure is speed. Absent extraordinary circumstances, a panel is required to hand down a decision within fourteen days of its appointment. As a practical matter, most proceedings are resolved within ninety days of commencement.
         
         
      8. Due to the limited opportunity to present evidence, the ability of a registrant to derail the proceedings by filing a lawsuit, and the limited relief available, most parties find ICANN dispute resolution to be unsuitable for complex or hotly-contested issues. Where, however, a trademark owner cannot locate the registrant or is unsure that it will even respond, ICANN dispute resolution can be an effective tool for quickly redressing abusive web sites.
         
         
4. Keyword Searching And Metatags
   
   

   The powerful search engines which drive the internet provide another context in which a trademark is stripped to a bare collection of letters, with no commercial context to mitigate likelihood of confusion. As with domain names, the courts have by and large attempted to apply traditional trademark concepts to alleged abuses, and, as with domain names, the courts have generally found a way to stop those who are perceived to be unfairly benefiting from the goodwill and recognition of another’s mark.
   
   

   1. The Technology and The Techniques
      
      
      1. Keywords are terms used in search engines to generate lists of domain names. The order in which those domain names are displayed obviously has a significant impact on how likely the viewer is to visit a site. Most typical searches today generate hundreds or even thousands of sites, and those outside the first ten or so are hugely unlikely to receive any attention at all. As a result, some search engines generate fees by selling the right to have a particular site show up first, or in the first few, when a particular keyword is searched. Keywords are also frequently used as triggers for banner ads which appear above or near the list of results.
         
         
      2. Metatags are a different technological approach to achieving the same result. They are terms which are embedded in the HTML header for a web page or site, which search engines use to compare against search terms which have been entered by the user. The tags are not visible in normal viewing of the site, but can affect how early and how often a particular site comes up on various search engines when a search term is entered.
         
         
   2. The Trademark Issues
      
      
      1. Difficulties arise, of course, when a company uses a keyword or metatag which embodies the trademarks of another. In Playboy Enterprises, Inc. v. Welles, 7 F. Supp. 2d 1098 (S.D. Cal.), aff’d mem. 162 F.3d 1169 (9^th Cir. 1998), the defendant was a former Playboy Playmate of the Year, and she promoted her personal website by advertising that fact, as well as by embedding the terms "playmate" and "playboy" as metatags in her site. The court refused to enjoin this practice because the terms were being used fairly to describe the defendant and her background.
         
         
      2. The Ninth Circuit distinguished Welles in Brookfield Communications, Inc. v. West Coast Entertainment Corp., 174 F.3d 1036 (9th Cir.1999). There, the defendant had embedded metatags of the term "MovieBuff" in its film website, a use which was challenged based on the plaintiff’s MOVIEBUFF trademark for competitive goods and services. The defendant argued that, as in Welles, the metatag it used had a descriptive meaning. The court disagreed, pointing out that the descriptive term is "movie buff," and defendant’s metatags lacked the all-important space in the middle: "[Defendant] can legitimately use an appropriate descriptive term in its metatags. But ‘MovieBuff’ is not such a descriptive term. Even though it differs from ‘Movie Buff’ by only a single space, that difference is pivotal. The term ‘Movie Buff’ is a descriptive term, which is routinely used in the English language to describe a movie devotee. ‘MovieBuff’ is not."
         
         
      3. Other badges of a lack of good faith descriptive use were catalogued in SNA, Inc. v. Array, 51 F.Supp.2d 554 (E.D. Pa. 1999). There, plaintiff was the manufacturer of a hobbyist amphibious aircraft sold in do-it-yourself kits called the Seawind, and defendants were in the business of selling engines for use in Seawinds, assembling kits for users, and publishing a newsletter for Seawind enthusiasts. Defendants’ websites embedded the term SEAWIND as a metatag, and the court held that the use was not a fair descriptive use, explaining: "Here, based on the repetitious usage and the evidence of defendants' general intent to harm plaintiffs, the court cannot find that this use is a similar good faith effort simply to index the content of the website; instead, it is a bad faith effort to confuse internet users that is likely to succeed. Defendants' meta tagging will thus be enjoined."
         
         
      4. Playboy’s unfortunate result in Welles was repeated in Playboy Enterprises, Inc. v. Netscape Communications, Inc., 55 F. Supp. 2d 1070 (C.D. Cal. 1999), and for similar reasons. The defendant owner of the Excite web portal had sold several hundred keywords to third party advertisers, among them "playboy" and "playmate." The court refused relief, stating that the terms were ordinary English language words that do not necessarily suggest sponsorship or endorsement by a particular company.
         
         
      5. Playboy doesn’t always lose, however. See Playboy Enterprises, Inc. v. Calvin Designer Label, 985 F. Supp. 1220 (N.D. Cal. 1997) (enjoining "playboy" metatags without discussion of the merits).
         
         
5. Linking And Framing
   
   
   1. The Technology and The Techniques
      
      
      1. Linking is the essence of the internet and its commercial embodiment, the world wide web. A link is a screen object which, when clicked by the user, loads a specific URL for another page, which can be located anywhere on the web. Particularly troublesome to some web site owners is the practice of "deep linking," in which the owner of a web site creates a link to a page several levels deep in the web site of another. The "deep" page may not have the same indicia of sponsorship as a home page, because the sponsor may have assumed that a user would only get to it by viewing the home page first. In some circumstances, the user may not even realize that the page viewed via the link is not part of the initial website. Even more important as a practical matter, though perhaps not as a trademark matter, the deep link may cause users to by-pass revenue generating advertising on the home page of the linked site.
         
         
      2. Framing is the practice of creating a link which causes the user to view another’s web page, but with a "frame" around the page created by the first sponsor. Thus, a link from company A’s website could cause users to view a page of company B’s website, but framed with company A’s marks, logo, trade dress, or other indicia of source. The potential for confusion as to source or sponsorship is obvious.
         
         
   2. The Trademark Issues
      
      
      1. Perhaps surprisingly, there has as yet been relatively little decided case law concerning the trademark issues raised by linking, and almost none concerning framing in the trademark context. One of the most notorious cases was settled before any decision on the merits was generated. In Ticketmaster Corp. v. Microsoft Corp., No. 97-3055 DDP (C.D. Cal., complaint filed April 28, 1997), Ticketmaster attacked Microsoft’s practices in connection with its "Seattle Sidewalk" city directory site. Among other practices, Ticketmaster alleged that Microsoft included a page within its site for users to be able to purchase event tickets, but instead of handling the transactions itself, users were linked without permission to the order page of Ticketmaster’s site. Ticketmaster claimed that the practice "misdescribed" Ticketmaster’s services, and enabled Microsoft to utilize Ticketmaster’s reputation to sell advertising. The case was reportedly settled in 1999 by Microsoft agreeing not stop deep linking to Ticketmaster’s site.
         
         
      2. Another settled case, Washington Post Co. v. TotalNews, Inc., Case No. 97 Civ. 1190 (PKL) (S.D.N.Y., complaint filed February 20, 1997), complained of defendants’ "totalnews.com" website, which displayed pages from plaintiffs’ websites carrying current news stories in frames designed to look like a television set, accompanied by defendants’ advertising. The case was reportedly settled by defendants agreeing not to frame, but remaining free to link to, plaintiffs’ sites.
         
         
      3. As the outcome of the above two cases suggests, the issue is whether the linking or framing is done in such a way as to confuse or mislead users as to whose content they are viewing, and from that point of view, framing is likely to be much more problematic than framing.
         
         
      4. In general, the basic concept of linking is unlikely to be viewed as troublesome unless the specific circumstances suggest that it is being done in a misleading way. In Bally Total Fitness Holding Corp. v. Faber, 29 F. Supp. 2d 1161 (C.D. Cal. 1998), the defendant maintained a gripe site targeted at the plaintiff called "Bally Sucks," and linked from that site to a pornographic site. The court rejected the idea that the link could be considered tarnishment, saying: "The essence of the Internet is that sites are connected to facilitate access to information. Including linked sites as grounds for finding commercial use or dilution would extend the [dilution] statute far beyond its intended purpose of protecting trademark owners from uses that have the effect of ‘lessening . . . the capacity of a famous mark to identify and distinguish goods or services.’"
         
         
      5. The idea in Bally Total Fitness that linking can create liability to the party linked from, by associating it with the party linked to, is an interesting twist which has arisen in some other cases. In Avery Dennison Corp. v. Sumpton, 189 F.3d 868 (9^th Cir. 1999), one of plaintiff’s claims was that defendant included domain names incorporating plaintiff’s marks on a website which also linked to a pornographic site, causing tarnishment. The court said that moving from the sites allegedly associated with plaintiff to the lewd site required linking through defendant’s home page, "which might remove any association with the ‘Avery’ and ‘Dennison’ trademarks that the Internet user might have had."
         
         
      6. The other end of the spectrum is illustrated by Archdiocese of St. Louis v. Internet Entertainment Group, Inc., 34 F. Supp. 2d 1145 (E.D. Mo. 1999), where the defendant’s unauthorized "papalvisit.com" and "papalvisit1999.com" sites featured links to adult websites. Not surprisingly, neither the Pope nor the court were particularly charitable toward the defendant.
         
         
      7. Another somewhat instructive case is Hard Rock Cafe International (USA) Inc. v. Morton, which was the subject of two unreported decisions by the Southern District of New York in 1999. The opinions’ utility is limited, however, by the fact that a license agreement was in place between the plaintiff and defendant, so the issue was not so much whether defendant’s linking and framing violated the trademark laws, but whether it violated the terms of the license. Nonetheless, it is interesting that the court held (1999 WL 701388) that when defendants linked and framed from their website displaying plaintiff’s licensed marks to third-party sites, there was no harm to plaintiff’s marks because defendants did not receive compensation from the linking: "Because any sales are made by third parties and not by defendants, and defendants do not receive any commission or remuneration on such sales, there is no basis for finding that this use of the Hard Rock Marks by defendants constitutes use of the Hard Rock Hotel Marks for the sale of merchandise or a function other than the promotion, advertising or marketing and broadcasting of defendants' Hard Rock Hotel."
         
         
      8. In a later opinion, however (1999 WL 717995), the court seemed to rethink that proposition, at least with regard to framing. It responded to defendants’ argument that the links were "merely technical connections between two independent sources of content" as follows: "Whether or not this is true with respect to hyperlinks, it is not true with respect to framing. . . . Framing is far more than a ‘technical connection between two independent sources of material.’ Through framing, the Hard Rock Hotel Mark and the Tunes site are combined together into a single visual presentation and the Hard Rock Hotel Mark is used to promote the sale of CDs by Tunes. Because the Tunes material appears as a window within the original linking page, it is not clear to the computer user that she or he has left the Hard Rock Hotel web site. . . . In light of this seamless presentation of the Tunes web page within the Hard Rock Hotel web site, the only possible conclusion is that the Hard Rock Hotel Mark is used or exploited to advertise and sell CDs." The framing was enjoined.