Interconnectivity

By William Fisher, Dequan Wang, Shel Woodruff, John Palfrey and Agnes Li

Table of Contents

Introduction
Case Study
Readings
Discussion Topics
Additional Resources

Introduction

The Internet is a vast network that connects many smaller groups of linked computer networks, on and through which information is stored and transmitted.  The "interconnected" character of the Internet is one of the things that makes it so popular and powerful in facilitating communication and electronic commerce. However, "interconnectivity" has also given rise to increasing legal controversy and turmoil. Various methods of enhancing or exploiting the web-like structure of the Internet have been attacked as violations of others' intellectual property or other proprietary rights.  In this module, we will consider the current status of those controversies.  The Case Study explores in some detail one recent case of this general sort.  The Readings then examine the histories of some related dimensions of "interconnectivity."

Case Study

Ditto.com operates a "visual search engine" on the Internet. Like other Internet search engines, it allows a user to obtain a list of related Websites in response to a search query entered by the user. Unlike the first-generation Internet search engines, Ditto retrieves images instead of text.  More specifically, it generates a list of reduced, "thumbnail" versions of photographs related to a user's query.  To see how the system currently works, please visit www.ditto.com. Enter a phrase in the search box and click "Search."  Then click on any of the "thumbnail" pictures retrieved by your query.

To compile its database of approximately two million thumbnail images, Ditto uses a "spider" or "crawler."  A spider is simply an automated computer program that explores (crawls) the Web, looking for information. The most common kinds of spiders are the ones that collect web pages for search engines to index.  (To learn more about this technology, see how Search Engines and Spiders work.)  When Ditto's spider finds an image, it makes a temporary, full-sized copy of it on Ditto's servers, then uses that copy to create a smaller, lower resolution "thumbnail" version.  Once the thumbnail has been prepared, the full-sized image is deleted.

As you have noticed if you have visited Ditto's current site, a user of the system may click on any picture to go directly to it on the web.  When a user clicks on a picture, a new browser window opens, displaying the website from which the image was originally taken and showing the image in its intended context.  An older, now superceded version of the service, operated by Arriba Soft (the predecessor company to Ditto), worked significantly differently, employing a process popularly known as "inline linking."  When a user of the original Arriba system "clicked" on a thumbnail image, he or she was taken to an "Image Attributes" page, which showed a full-sized version of the image along with a textual description of the image, a link to the site from which it was taken, the Arriba banner, and Arriba advertising.  The full-sized image did not reside on Arriba's servers but was transmitted to the user’s browser directly from the originating page.  However, the typical user, seeing the image surrounded by text and advertising generated by Arriba, would assume that Arriba had copied it.

Leslie Kelly is a photographer who maintained two Web sites.  www.goldrush1849.com (currently off line but available at http://web.archive.org/) provides a "virtual tour" of California's gold rush country and promotes Kelly's book on the subject.  www.showmethegold.com markets corporate retreats in California's gold rush country. Approximately thirty-five of Kelly's images were indexed by the Arriba crawler and put in Arriba's image database. As a result, these images were made available in thumbnail form to users of Arriba's visual search engine.

Angered by the nonpermissive use of his photographs -- and, in particular, by the inline linking system through which they were displayed to users -- Kelly sued Arriba for copyright infringement.  (An account of the origins of the case, colored by strong support for Kelly's position, can be found at http://netcopyrightlaw.com/kellyvarribasoft.asp.)

Before reading the responses of the various courts that have heard the case, try to formulate your own opinion of the proper outcome.  In doing so, consider the following questions:
 

1.  All versions of the defendant's search engine (including the present one) assist web users in finding images and then downloading them to the users' own hard drives -- for their own enjoyment or for commercial purposes.  Unauthorized copying of that sort may itself violate the copyright statute.  Should the defendant be liable for facilitating such behavior?

2.  Arriba's first and second-generation systems enabled users to "deep link" directly to the pages containing retrieved images, and thereby bypass the "front page" of the originating Web site.  As a result, these users would be less likely to view all of the advertisements on an originating Web site or to view the Web site's entire promotional message.  Is harm of this sort sufficient to render the systems illegal?

3.  One of the judges who considered the case reasoned that "where, as here, a new use and new technology are evolving, the broad transformative purpose of the use weighs more heavily than the inevitable flaws in its early stages of development." Do you agree?

4.  A robots.txt file is a widely recognized standard used to stop search engines from crawling a web site. It works as a de facto automatic copyright notice for spiders.  Is it fair to infer that websites not employing this device are "inviting" or "acquiescing" in visits by spiders?  If your answer is “yes”, what rights, if any, do you think the website has given up by not including a robots.txt file?

So how was the case actually resolved?  At the trial level, Arriba prevailed.  Judge Taylor of the Central District of California ruled that, although Arriba, by displaying without permission thumbnail versions of Kelly's photographs, had violated section 106 of the Copyright Act, its conduct was nevertheless excused by the "fair use" doctrine embodied in section 107.  Kelly v. Arriba Soft Corp., 77 F. Supp. 2d 1116, 1121 (C.D. Cal. 1999) available at 160 F.Supp.2d 266 (D. Conn. 2001), available at http://www.ctd.uscourts.gov/Opinions/080701.GLG.PrimePub.pdf.  In February 2002[MSOffice1] , a panel of the Court of Appeals for the Ninth Circuit reversed.  Kelly v. Arriba Soft Corp., 336 F.3d 811 (9th Cir. 2003).  The panel agreed with Judge Taylor's analysis concerning the legitimacy of Arriba's creation and use of the thumbnail versions of Kelly's images under fair use.  Id at 822.  The Court then held that the fair use doctrine did not excuse the inline linking to the full-sized versions of Kelly's images by Arriba's original search engine, and then remanded back to the lower court on whether this would infringe Kelley’s copyright, so this question is still unsettled.  Id.

Readings

An Overview

The legal and technical issues involved in Web site publishing are intertwined. To see how they fit together, please read the following survey:  Martin J. Elgison and James M. Jordan III, Trademark Cases Arise from Meta-Tags, Frames: Disputes Involve Search-Engine Indexes, Web Sites Within Web Sites, As Well As Hyperlinking, at http://cyber.law.harvard.edu/property00/metatags/mixed1.html  (1998).  Some of the individual issues discussed by Elgison and Jordan, and developed by many other authors on topics related to copyright and other doctrines, are explored in more detail below:

(1) Linking

"Linking" (also referred to as “hyperlinking” or “hypertext linking”) is the primary mechanism by which sites on the World Wide Web are connected. By simply clicking on a "live" word or image in one Web page, the user can view another Web page located on server elsewhere in the world, or simply elsewhere on the same server as the original page. This system helps gives the Web its extraordinary communicative power and linking to another site without prior permission has become standard practice on the Internet.  Two legal theories, fair use and implied license, have been advanced for a right to link to a copyrighted web site but the law on linking is still unsettled.  See Carl S. Kaplan, Is Linking Always Legal? The Experts Aren't Sure, Apr. 6, 1999, at http://www.nytimes.com/library/tech/99/08/cyber/cyberlaw/06law.html. Under some circumstances, a link may undermine the rights or interests of the "target" site. 

(a) Trademarks in Hyperlink Uniform Resource Locators (URL’s)

Every hyperlink consists of two parts, the text or image displayed to the user and the URL (i.e. the Internet address) of the web page to display when the hyperlink is selected.  URL’s normally include the owner’s domain name, which often includes trademarked terms (e.g. “http://ford.com”).  Can a trademark holder prevent a third party from including a hyperlink to their web site by claiming that the use of the URL is a trademark violation?

In a case directly addressing this issue, Ford Motor Co. v. 2600 Enterprises, 177 F. Supp. 2d 661 (E.D.Mich. 2001), a Federal Judge refused to enjoin the defendant from linking to Ford’s web site using a disparaging hyperlink.  The judge wrote,

This court does not believe that Congress intended the FTDA [Federal Trademark Dilution Act] to be used by trademark holders as a tool for eliminating Internet links that, in the trademark holder's subjective view, somehow disparage its trademark.  Trademark law does not permit Plaintiff to enjoin persons from linking to its homepage simply because it does not like the domain name or other content of the linking webpage.

Id at 664.  However, the opinion is limited because it relies strongly on the non-commercial nature of the defendant’s use of the mark.

In 1996 the State of Georgia passed a state statute that, inter alia, criminalized the transmission of any data that uses a registered trademark through a computer network without the owner’s permission.  A Federal Court subsequently issued a permanent injunction against enforcement of the statute, A.C.L.U. v. Miller, 977 F. Supp. 1228 (N.D. Ga. 1997).  The plaintiffs brought the case primarily because the statute also banned anonymous speech on the Internet, but in ruling that the statute was excessively broad the court noted that enforcing the law would prohibit the use of web links because they often contain trademarks or trade names.

Bosley???

Neither of these cases definitively settles the question of whether the inclusion of another’s trademark within a hyperlink on a commercial web site is justified under either a fair use or an implied license defense.  However, both courts noted that non-permissive linking is a necessary requirement for the World Wide Web as it currently exists. 

(b) Avoiding Advertisements or Message by Deep Linking

Suppose, for example, that X sets up a homepage for her site. On the homepage she places some advertisements, from which she hopes to make some money. The homepage also contains links to various internal subordinate pages, which contain content that X believes consumers wish to see. Y then creates his own Web site, which contains links to X's subordinate pages. The net result is that visitors to Y's site will be able to gain access to X's material, without first viewing X's homepage where the advertisements are placed.  This activity is called "deep linking."

One of the first major cases in the law of linking arose in Scotland. At issue was the use by one news Web site of the headlines of another news Web site to link to the latter's stories. Read about this case in the following article: Scottish Court Orders Online Newspaper to Remove Links to Competitor's Website.

A little over two years after the Shetland News decision, a California court was presented with an analogous case. In reading about this dispute, consider the factual, legal, and cultural differences that might have led to its drastically different conclusion: CA Judge Dismisses Copyright Claims Based on Linking. (explaining Bernstein v. J.C. Penney, Inc., No. 98-2958-R (CD CA, dismissal Sept. 22, 1998)).

(c) Linking to Unlawful Material

Recently, a host of new cases involving linking have arisen. The most interesting of these involve allegations of "contributory copyright infringement." On this theory, it is asserted that Web sites should be held accountable when they provide links that assist Web users to engage in nonpermissive copying of copyrighted works.  The first case took place in Holland and involved a copyright claim regarding the Church of Scientology, which has since been involved in a dispute with Google over similar matters.   Another case arose in Utah. See Intellectual Reserve, Inc. v. Utah Lighthouse Ministry, Summary (D. Utah 1999), available at http://eon.law.harvard.edu/property00/metatags/cases.html#ULM.

In both Intellectual Reserve and Universal City Studios, Inc. v. Reimerdes, 111 F.Supp.2d 294 (S.D.N.Y. 2000), the defendants originally posted without permission copyrighted material belonging to the plaintiffs.  In both cases, after Federal Courts enjoined the defendants to remove the offending material, the defendants removed the material but posted directions on how to obtain the material on other web sites, along with links to the material on other sites.  In both cases the defendants also chose to taunt the court by including claims that the injunction was not going to stop them from making the copyrighted material available.  See e.g. Reimerdes, F.Supp.2d at 294.

In the Intellectual Reserve case, the Court found that the defendant’s actions were likely to constitute contributory copyright infringement and issued an order not to post “addresses to websites that defendants know, or have reason to know, contain the material alleged to infringe plaintiff's copyright.”  available at http://www.techlawjournal.com/courts/lds/19991206.htm.

The district court in Reimerdes also prohibited the defendant from posting links to the offending material.  The Second Circuit Court of Appeals in Universal City Studios, Inc v. Corley,  upheld the lower court decision by carefully distinguished a hyperlink as having both an “action component” (the ability to go to another web site by clicking on it) and a speech component (the text of the URL itself).  273 F.3d 429 (2d Cir. 2001).   The Second Circuit noted that the “linking prohibition is justified solely by the functional capability of the hyperlink,” id at 456, suggesting that the actual URL text was protected speech.

In Japan, if a website links to an illegal site, such as pornography, the web site owner could be indicted for having "aided and abetted" the public exhibition of illegal materials. (citation?)

(d) Avoiding Agree Pages by Deep Linking

Another issue arises when X sets up a homepage with an “agree” link to prevent visitors from entering X’s site without agreeing to various terms. Visitors to X’s page must click on “agree” to secure that they will follow the terms of X’s homepage before going on to subordinate pages. Y then creates a Web site that links to X’s subordinate pages. The result is that the visitors to Y’s site will be able to gain access to X’s materials without promising to agree to X’s terms.

The case that most squarely presents the legal issues raised by this practice is Ticketmaster Corp. v. Tickets.com, Inc., 54 U.S.P.Q.2d 1344 (C.D. Cal. 2000). Ticketmaster sought to enjoin Tickets.com from linking to subordinate pages of ticketmaster.com to buy tickets to concerts and other events. The link from Tickets.com bypassed the Ticketmaster homepage, which listed “terms and conditions,” one of which prohibited deep linking. Ticketmaster filed, among other complaints, a breach of contract claim, on the ground that Tickets.com had circumvented the “terms and conditions” on the homepage. Ticketmaster argued that this case was analogous to a "shrink-wrap license" case, where the packaging on the outside of a product states certain terms and conditions to which purchasers, once they open the package, are bound.  [The conflicting doctrines pertaining to this marketing stratagem were reviewed in the second module.]  Ticketmaster claims that Tickets.com in effect had allowed users to bypass the packaging of the Ticketmaster site. Id.

The US District Court in California disagreed and held that this application of deep linking was permissible.  Ticketmaster Corp. v. Tickets.Com, Inc., 2000 U.S. Dist. LEXIS 12987 (D. Cal. 2000).  In denying the plaintiff a preliminary injunction, the court emphasized the fact that, unlike many web sites that make the visitor click on "agree" to the terms and conditions before going on, ticketmaster.com did not require any manifestation of assent.  The court distinguished this case from “shrink-wrap license” cases, because whereas the packaging on a product is exposed and noticeable, the “terms and conditions” on the Ticketmaster site can be overlooked since there is no “agree” page. Id. at *8.  The court further pointed out that the terms and conditions on the Ticketmaster site are set forth so that the user needs to scroll down the home page to find and read them. Hence, the court stated, users can easily go on to subordinate pages without reading the terms and conditions. The court concluded that mere inclusion of terms and conditions on a web site does not create a contract with anyone using the site, because there are no facts showing either knowledge of the terms or agreement to them.  Id.  A panel of the Court of Appeals for the Ninth Circuit summarily affirmed. Ticketmaster Corp v. Tickets.com, Inc., 2 Fed. Appx. 741 (9th Cir. 2001).

Since this decision, Ticketmaster has employed technical methods of blocking direct access to subordinate pages, so that any link from an outside site takes the user to the main page of ticketmaster.com. cite???  It would appear from this ruling that deep linking that does not circumvent “agree” pages would be free from liability. Ticketmaster Corp, 2000 U.S. Dist. LEXIS  12987 at *8.  However, this decision was made in a minute order not intended to have precedential authority.  Hence, future decisions may be necessary to determine whether deep-linking that does not bypass an agree page is permissible.

(2) Framing

The related practice of "framing" may also serve to undermine the rights of Web site owners.  The use of "frames" allows a Web page creator to divide a user's Web browser window into several separate areas.  The programmer of the Web page can dictate what goes into each frame. Commonly, a Web site designer creates a page that at all times displays one frame containing the name of the Web site and other identifying information.  For example, a Web site employing frames might always show the original Web site's graphic logo on the top of the page while allowing the user to view the "framed" site in a different frame. (For an illustration, please visit www.totalnews.com.   Click on "Local", then "Birmingham Post-Herald."  The latter site, you will see, is "framed" within the TotalNews web site.)

Persons and organizations whose sites are "framed" without permission have objected to this practice on three main grounds:  First, they argue that Web surfers might easily be confused concerning the relationship between the framed site and the framing site. Second, the framing site is parasitic, unfairly deriving traffic and, often, offering advertising to users visiting the legally protected work on the framed site.  Third, the activity of framing distorts the aesthetic of the target site.

One of the first legal disputes arising out of objections of these sorts involved Total News (http://www.totalnews.com/), a Web site that links to several major news sites. Although the case brought against Total News was settled (and therefore never resulted in any substantive law), its facts provide an excellent illustration of the difficulties presented by framing. Please read the attached excerpt from Mary M. Luria, "Controlling Web Advertising: Spamming, Linking, Framing, and Privacy."

Another framing case in which the defendant prevailed is Futuredontics, Inc. v. Applied Anagramics, Inc., 1998 U.S. App. LEXIS 17012 (9th Cir. Cal. July 23, 1998), available at http://cyber.law.harvard.edu/ilaw/property/Futuredontics.htm. In retrospect, the outcome probably derived from a crucial omission in the plaintiff's case: no evidence of damage from the defendant's activities had been shown.

The most salient recent case to have addressed the framing issue was Arriba Soft, 77 F. Supp. 2d at 1116, discussed in the Case Study, above.  With the withdrawal by the Ninth Circuit of their holding that Arriba's original search engine impermissibly "displayed" the plaintiff's copyrighted photographs, there is little case law to establish the legal limits of framing third party’s website.

(3) Meta Tags

The pages (“source code”) of most Web sites are primarily written in the HyperText Markup Language (HTML).See, e.g., the W3C’s page regarding HTML.  This language consists of a set of "tags" that can be used to format and arrange text, images, and other multimedia files.  (The HTML source code for a web page may be viewed in Microsoft’s Internet Explorer by right clicking and selecting “View Source”.)  "Meta tags" are text that have no visible effect on the Web page.  Instead, they exist in the source code for a Web page to assist search engines in ascertaining the content of the page.  Problems arise when website designers include in their lists of meta tags the names or descriptions of other companies.

(a) Trademarks in Meta Tags

Suppose, for example, that Coca Cola used the keyword "Pepsi" in its meta tags. Web surfers who used search engines to obtain information about "Pepsi" would then be directed to Coca Cola's Web site. The gravitational power of meta tags has been diminishing as search engines have become more sophisticated in the ways they strive to ascertain the contents of Web sites.  The search method of google.com, which carries out more than half of the searches on the Web as of 2004, known as “PageRank™,” is less reliant on meta tags than previous search technologies.  Nevertheless, meta tags retain sufficient capacity to shape behavior on the Internet to continue to generate legal controversies.

One of the first such cases was Oppedahl & Larson v. Advanced Concepts, Civ. No. 97-Z-1592 (D.C. Colo., July 23, 1997). The defendants were found to have no defensible reason for using the name of Oppedahl & Larson (a law firm that had dealt in domain name disputes) in their meta tags.  By using the plaintiff's name in the tag, the defendants were trying to capture traffic that would gain them domain name registration fees or web site hosting clients. The court barred them from using the name without authorization. For more details regarding this case, please read David J. Loundy, "Hidden Code Sparks High-Profile Lawsuit."

By contrast, in Playboy vs. Terri Welles, District Judge Judith N. Keep refused to grant an injunction against Terri Welles, Playboy's 1981 Playmate of the Year, to prevent her from using terms such as "Playmate" and "Playboy" on her web pages and within her meta tags. The judge ruled that the defendant was making fair use of Playboy's marks because she was using them to describe herself truthfully on the web site.  7 F. Supp. 2d 1098 (D. Cal., 1998).  The Ninth Circuit Court of Appeals affirmed the district court’s ruling (with the exception of Welles’s use of the abbreviation “PMOY” as a watermark on the wallpaper of her site) on Feb. 1, 2002.  Playboy Enters. v. Welles, 279 F.3d 796 (9th Cir., 2002), available at http://caselaw.lp.findlaw.com/data2/circs/9th/0055009p.pdf.  For more details concerning the case, see Playboy Enterprises v. Welles.

A recent decision in a similar vein is Bihari v. Gross, 119 F. Supp. 2d 309 (U.S. Dist., 2000).  Plaintiffs Marianne Bihari and Bihari Interiors, Inc., owner of the service mark "Bihari Interiors," brought a trademark infringement action against Gross, operator of a website that harshly criticized Bihari's business practices.  Among the many bases of Bihari's complaint was Gross' use of the term "Bihari" in his meta tags.  In finding for the defendant, the district court emphasized the fact that Bihari did not have a website of her own.  Thus, users were unlikely to experience even "initial interest confusion" when searching the Internet for information about Bihari Interiors. Citing Playboy v. Welles, the court further held that Gross had used the terms "Bihari Interiors" and "Bihari," not as trademarks, but rather in a descriptive sense to fairly identify the content of his websites.  For more information on this case, see Bihari v. Gross.

An entirely different tone was set by Judge Spero of the Northern District of California in a case involving two rival software companies.  Oyster Software Inc. v. Forms Processing Inc., No. C00-0724 JCS, 2001 US Dist. Lexis 22520 (N.D.Cal. November 13, 2001).  A consultant employed by Forms Processing Inc. (FPI) to "increase its prominence with search engines" allegedly copied some of the metatags in Oyster Software's website and then included them in FPI's website.  Despite the fact that FPI, upon being notified by Oyster, promptly removed the offending material, Oyster brought suit alleging copyright infringement, trademark infringement, and "trespass to chattels." Judge Spero rejected most aspects of FPI's motion to dismiss the suit.  Id. available at http://www.phillipsnizer.com/library/cases/lib_case299.cfm.  The most striking aspect of his ruling is his determination that, merely by visiting Oyster's website in order to obtain the metatags in question, FPI may have been guilty of "intentionally interfering with the possession of personal property" (namely Oyster's servers).  Id. at *36-37.  This represents a potentially radical extension of the theory of "trespass to chattels," first introduced in the Ebay case (discussed below).

The newest decision in this category involves efforts by Victoria's Secret to defend their famous trademark for lingerie and women's apparel against encroachment.  Victoria's Secret Stores v. Artco Equip. Co., 194 F. Supp. 2d 704 (D. Ohio, 2002).  One of the many ways in which Artco Equipment allegedly misused that mark was by employing the term, "victoria's secret," in the metatags for websites in which Artco offered lingerie and sex toys.  In March 2002, Judge Smith of the Southern District of Ohio brushed aside all of the defendant's effort to excuse this conduct, concluding that the case presented classic examples of "bait and switch" use of a trademark, "initial interest confusion," and trademark dilution.  Id.

(b) Generic Words

What happens when the word used in a metatag is an everyday term? Suppose, for example, that Coca Cola sought to enjoin Pepsi from using “always” in a meta-tag on the web site, on the ground that it created confusion with Coca Cola's trademarked slogan, “Always Coca Cola."  One might think that such a common term, taken out of context, would be free for use by anyone.  But a recent decision by a federal court in Connecticut casts doubt on that assumption.  In Prime Publishers, Inc. v. American-Republican, Inc., Prime sought to enjoin American from using the word “voices” in the meta-tag of the American site based on the Anticybersquatting Consumer Protection Act and related claims.  160 F.Supp.2d 266 (D. Conn. 2001), available at http://www.ctd.uscourts.gov/Opinions/080701.GLG.PrimePub.pdf.  The court agreed with Prime, because it found that although “voices” is a generic word, there was no need for American to use the word "voices" to describe the features or qualities of its site. Prime owned a newspaper called “Voices” and American owned a newspaper called “the Waterbury Republican-American.”  The court reasoned, to members of the newspaper world,  the word “voices” suggests the Prime paper. Thus a likelihood exists that business advertisers and internet users could be confused as to the source, sponsorship, affiliation, or endorsement of the website.  No proof of actual confusion was necessary; a risk of such confusion was sufficient. The court prohibited American from using the word "voices" in any part of the web site including the meta-tags.  Prime Publishers has since been cited in subsequent cases that also involved the Anticybersquatting Consumer Protection Act (ACPA).

 

(4) Search Engines

(a) Finding Queries

Search Engines are popular web-retrieval tools that match a search query submitted by an Internet user with the websites whose content best corresponds to the submitted search terms.  A few search engines sell to websites the right to be placed high on the lists generated in response to users' queries.  Most search engines, however, create their listings automatically. While the precise methodology varies from search engine to search engine, this listings creation process is usually achieved as follows:  The search engine constructs and activates a software program called a "spider," also known as a "crawler" or "robot."  The spider visits web pages, reads them, follows the embedded links to other pages, etc.  Everything the spider finds is deposited in the search engine's index.  Finally, the search engine's software sifts through the millions of pages recorded in the index to find matches to a given query and then ranks those matches in order of relevance.  For general information about search engines, please refer to http://searchenginewatch.com.

As the Case Study suggested, this activity is especially likely to trigger accusations of copyright infringement when used by visual search engines, such as Ditto.com.  (For a historical and technical perspective on visual search engines, please take a look at AltaVista's Photo Finder service, introduced in 1998.)  In that cases, the defendant’s primary line of defense is likely to be the fair-use doctrine, similar to the case involving using thumbnail pictures when users are searching for pictures.[1]  It is possible that, in the future, similar challenges could be brought against search engines that prepare brief textual summaries, typically copied from the listed websites.  (See, for example, each entry after a Google search, which lists both text [AL2] and thumbnail images[2], as well as aggregated news feeds and other content.)

(b) Posting Banner Ads

Some search engines use keywords to trigger banner ads at the top of their search result lists. Netscape, Excite, and Yahoo, for example, have long been selling certain words to advertisers who want their banner ads to appear when those words are entered by an Internet user.  Recently, these search engines have also begun selling trademarked terms as keywords. The latter practice has given rise to litigation.

In Playboy Enters. v. Netscape Communs. Corp., 354 F.3d 1020 (9th Cir. 2004), Netscape and Excite sold the terms "playboy" and "playmate" as keyword search terms prompting banner ads for adult entertainment sites. Playboy, which owns the trademarks in both terms, claimed that this practice would produce both consumer confusion and the "dilution" of its marks.  The United States District Court for the Central District of California disagreed and granted summary judgment to the defendants.  Playboy Enters. v. Netscape Communs. Corp., 55 F. Supp. 2d 1070 (C.D. Cal. 1999).  The Ninth Circuit Court of Appeals reversed, finding a possibility of initial consumer confusion and trademark dilution, and remanded the case back to the District Court.  Playboy Enters., 354 F.2d at 1022.  “Initial interest confusion” occurs when a consumer is initially tricked into visiting a web site unaffiliated with the trademark.  Even though the consumer immediately realizes that the trademark holder does not sponsor the new site, the site offers content close enough to what he or she was looking for that the consumer never returns to their search for the original trademarked site.  Id. at 1024-25.  Under this theory, the Court suggested that if the banner ads were clearly labeled as not belonging to Playboy Enterprises, they might not be infringing.

In a similar case, Estée Lauder, Inc. filed lawsuits against Excite and an online cosmetic company, Fragrance Counter, Inc., in the United States, Germany, and France. In each case, Estée Lauder attempted to enjoin Excite's sale of the search term, "Estée Lauder," to an online fragrance discounter as a banner ad trigger.  In Germany, a District Court agreed with the plaintiff -- and thus prohibited Excite and iBeauty, the successor to the Fragrance Counter, from using the keywords "", "Clinique", and "Origins", to trigger iBeauty’s banner ads on the Excite Website. The court also ordered Excite and iBeauty to cease the use of the trademarks within banner ads. The court reasoned that Excite and iBeauty had engaged in improper exploitation and unfair business practices under German law. In the parallel case in the United States, Estée Lauder’s motion was dismissed. Estee Lauter Inc. v. Fragrance Counter Inc., 99 Civ. 0382 (D.S.N.Y. 1999), available at http://www.nysd.uscourts.gov/courtweb/pdf/D02NYSC/99-05327.PDF. However, Estée Lauder and iBeauty ended up settling in an agreement in which iBeauty voluntarily agreed to refrain from using the Estée Lauder trademarks as Internet keywords. Afterwards, Estée Lauder and Excite settled in an agreement in which Estée Lauder’s ads will now pop up when users enter those words.

See Beth I. Z. Boland, Daniel B. Trinkle & Christine M. Baker, “Initial Interest Confusion” and the Use of Meta-Tags and Keyed Banner Ads in Internet Trademark Law, 45-OCT B. B.J. 6 (2001) for additional discussion of search-engines' use of trigger words for banner ads.

(5) Spiders

Occasionally, spiders are used by organizations other than search engines. Bidder's Edge (BE) was one such organization.  BE (now BidXS) operated an online "auction aggregation service."  In other words, it did not itself operate an Internet-based auction, but rather provided users links to auctions being held on other sites.  One of the sites to which BE provided such connections was eBay, the largest and most famous auction system in the world.  To gather and update information from Ebay, a BE spider "crawled" into the Ebay system approximately 100,000 times a day.  Ebay objected.  When negotiations broke down, Ebay brought suit and ultimately prevailed -- not on the basis of copyright or trademark infringement, but on the ground that the nonpermissive searches conducted by BE's spider constituted a "trespass to chattels."  2000 WL 1863564 (N.D.Cal. July 25, 2000), available at http://pub.bna.com/lw/21200.htm.  The case was appealed to the federal Court of Appeals for the Ninth Circuit.  Twenty-eight law professors submitted a brief to the Court of Appeals, arguing that the district court's trespass theory would unavoidably extend to linking. "Linking, like search engines, would now seem to exist at the sufferance of the link-to party, because a Web surfer who followed a 'disapproved' link would be trespassing on the plaintiff's server just as Bidder's Edge was." The professors believed that "requiring permission to link has dire implications for free speech as well as electronic commerce." (Please click here to read the entire brief of Amici Curiae.)  For better or worse, the court never took up this argument because in July 2001, on the eve of oral argument, Bidder's Edge, nearing financial and business collapse, settled the suit[AL3] .

For a discussion of the difficulty of applying old legal doctrines to new kinds of technology -- and a proposal concerning a sensible way of handling unwanted spiders, see Maureen A. O'Rourke, Shaping Competition on the Internet: Who Owns Product And Pricing Information?, 53 VNLR 1965  (2000). 

Discussion Topics

1.  Who should have won in the Ditto.com and Bidder's Edge cases?

2.  Are the social advantages of interconnectivity sufficiently strong that the plaintiffs in all of these cases should lose?  If not, in what ways should the law limit the ability of web site designers to make nonpermissive connections to other web sites?

3.  Some observers think that the unsatisfactory ways in which courts have thus far handled cases involving interconnectivity demonstrate, among other things, how poorly suited traditional intellectual property and tort doctrines, such as trespass to chattels, are for addressing conflicts that arise on the Internet.  In other words, these judicial decisions are sometimes cited in support of the general proposition that we need an entirely new legal doctrine to govern the Internet.  Do you agree?

4.  Web site owners who wish to avoid "being framed" by other sites can use various forms of coding, including relatively simple HTML code, to establish strong (though not perfect) defenses. That is, a Web page can be encoded so that most Web browsers will not allow it to appear within a frame, but only in its own, separate browser window.  (Click here to learn about these techniques.)  To what extent do you believe that Web site owners wishing not to be framed should be required to employ such technical solutions before they can seek legal redress?

*****

Additional Resources

Supplemental Readings

Law Primer: Copyright Basics: If you are unfamiliar with this body of law and would like to know more, click here to check out our copyright primer and related links.

Law Primer: Trademark Law. Much of the debate concerning the legitimacy of meta tags involves trademark law. If you are unfamiliar with this body of law and would like to know more, check out the trademark section in the library.

Mark Sableman, Link Law Revisited: Internet Linking Law At Five Years (2001) and the July 2002 Supplement (2002). A thorough examination of legal implications of hyperlink use on the Internet. A long but comprehensive look at the many legal arguments that can be made for and against users of hyperlinks.

Brad Templeton, 10 Big Myths about copyright explained. "An attempt to answer common myths about copyright seen on the net and cover issues related to copyright and USENET/Internet publication."

Timothy J. Walton, Copyright for Webmasters. This article gives a basic description of copyright law, aimed at those who are involved in the production of Web sites.

Timothy J. Walton, Trademark Law for Webmasters. Same as above, but focusing on trademark law and the Web.

Jeffrey R. Kuester and Peter A. Nieves, "What's all the hype about hyperlinking?": Short article advocating the activity of linking and asserting that it is "the essence of the Web". Also discusses key linking cases and the use of state common-law claims in the Internet context. Looks at the feasibility of applying state law to regulate Internet activities. Concludes by commenting on recent federal attempts to regulate the Internet through the now defunct Communications Decency Act (CDA).

Mary M. Luria, "Controlling Web Advertising: Spamming, Linking, Framing, and Privacy.": In-depth article, which focuses on "the delicate balance between protecting the rights of people and businesses and preserving the spirit of open communication that is the hallmark of the Web." Areas examined include spamming, linking, framing, and consumer privacy.

Matt Jackson, "Linking Copyright to Homepages.": Law review article discussing linking in detail as well as related doctrinal considerations. Provides background on constitutional basis and economic rationale for copyright law. Primer included on how the Internet works.

John M. Mrsich, Meeka Jun, "Terms You Need to Know: Search Engines.": Description of how search engines work; terms used; and distinctions among search engines, search directories, and search managers.

The Perkins Coie Internet Case Digest: excellent and diligently updated database of Internet case law.

The Link Controversy Page: A very comprehensive site providing links to articles and cases. Don't try to cover the whole site! Just skim it to get a sense of the debate.

Meta Tag Lawsuits: A concise overview of all meta tag cases and links to articles and other resources. Again, you shouldn't try to review the entire site, just skim the case overviews.

Linking and Liability: Nice overview of the linking controversy. Begins by describing, in detail, the activity of linking. Then looks at the linking issue against the backdrop of current legal doctrine such as derivative-work rights under copyright law, the doctrine of “passing off,” defamation, and trademark infringement. Concludes by providing a summary of the Shetland News case, which is discussed in context, above.

Cases

Ford Motor Co. v. 2600 Enterprises, 177 F. Supp. 2d 661 (E.D.Mich. 2001).  The court refuses to grant a preliminary injunction against the magazine 2600 for using Ford’s trademark as a URL (http://ford.com) in a disparaging hyperlink.

In A.C.L.U. v. Miller, 977 F. Supp. 1228 (N.D. Ga. 1997), prohibits enforcement of a Georgia state statute noting, in footnote 5, that requiring permission from trademark holder to use a mark in a computer network would “prohibit the current use of web page links.”

The Washington Post Co., et al. v TotalNews, Inc.,97 Civ. 1190 (S.D.N.Y., filed Feb. 2, 1997): Case settled June 5, 1997. No opinion was issued. You can see the Complaint or the text of the Settlement (Order of Dismissal).

Futuredontics, Inc. v. Applied Anagramics, Inc., No. 97-6991, 1998 U.S. Dist. LEXIS 2265 (C.D. Cal), aff’d, 1998 U.S. App. LEXIS 17012 (9th Cir. 1998) (unpublished opinion).  Court denies preliminary injunction against AAI's use of a framed link to Futuredontics' web site. The denial was upheld on appeal. For additional information see this article discussing the case. 

Shetland Times Ltd. v. Dr. Jonathan Willis and ZetnewsLtd., Scotland Court of Sessions (Oct. 24, 1996): Case settled Nov. 1997; no opinion issued.

Ticketmaster Corp. v. Microsoft Corp. [CV 97-3055RAP] (C.D. Cal., filed April 28, 1997): Case settled; no opinion issued. Text of Complaint.

Bernstein v. J.C. Penney, Inc., No. 98-2958-R (CD CA, dismissal Sept. 22, 1998). Plaintiff claimed copyright infringement against defendants who maintained links that eventually led site with allegedly infringing photographs.  Case dismissed. For a discussion of the case see this article.

Church of Scientology v. Dataweb et al, Case No. 96/1048 (Dist. Ct. of the Hague, Holland, June 9, 1999). Court found ISP liable for its customer's posting of infringing material on the Web. Liability was found because there was notice, which could not be reasonably doubted, of the poster's activities.

Oppedahl & Larson v. Advanced Concepts [Civil Action 97-3-1592 (D. Colo.) filed July 23, 1997]. You can see the complaint and the permanent injunctions entered with respect to the various defendants, namely Welch and Advanced Concepts, and Professional Website Development, Internet Business Services.

Playboy Enterprises Inc. v. Calvin Designer Label [DCN Calif, No. C 97-3204; 9/8/97]. You can see the Complaint and the Preliminary Injunction Order.

Playboy Enterprises, Inc. v. Welles No. 98-CV-0413-K (JFS) (S.D. Cal. Apr. 21, 1998), affirmed in part Playboy Enterprises, Inc. v. Terri Welles, et al., 162 F.3d 1169 (9th Cir. 2002).  Case discussion.  

Niton Corp. v. Radiation Monitoring Devices, Inc, 27 F. Supp. 2d 102 (D. Mass. 11/18/98): Court granted preliminary injunction against defendant, barring it from using anything in its meta tags that would mislead visitors in a number of different ways.

SNA Inc. v. Array et al, No. CIV A 97-7158 (E.D. Pa., June 9, 1999). Enjoined defendant from using plaintiff's name in its meta tags.

In Universal City Studios, Inc. v. Reimerdes, 111 F. Supp. 2d 294 (S.D.N.Y. 2000) aff’d sub nom. Universal City Studios, Inc v. Corley, 273 F.3d 429 (2^nd Cir. 2001) the court enjoined the defendants from linking to copyrighted material based on the “action” component of  the hyperlink (to avoid first amendment problems).  A complete set of materials on this case is available here.