Email	Print Reprint
	add to:
	Del.icio.us Digg Google Furl	Slashdot Y! MyWeb Blink

In the beginning, Tim Berners-Lee created HTML. But the Web was without form, and void. And Tim said, "Let there be links" and there were links. And Tim saw that the links were good.

In the beginning, we only had links. We didn't think about different kinds of links, and in the minds of most Web developers, that's still the case today. After all, HTML doesn't recognize any differences between links. They all take the same basic form: <a href="http://www.example.org">link</a>. Links are just, well, links.

Lawyers, however, have started to make some distinctions between links. In the legal world, some links are acceptable, some links are acceptable some of the time, and some links are prohibited outright. To any seasoned netizen, many of the distinctions the legal profession makes are utter rubbish. But if you think there is only one kind of link, read on. Courts are unpredictable, especially when dealing with technology, and it may be worth reviewing the current lay of the land.

Things have changed since the day Berners-Lee and friends invented the World Wide Web, and not for the better.

Deep Trouble

The Dallas Morning News has a site that's virtually indistinguishable from any other newspaper site on the Web. The site's "home page," much like the physical newspaper's front page, contains headlines for what's inside with links to the articles. There's one important difference on the paper's Web site, however. At the bottom of the DallasNews.com home page, you'll see a link to Terms. Click on that link and you'll be directed to the terms of service agreement. A portion of it states: "If you operate a Web site and wish to link to this site, you may link only to the home page of the site and not to any other page or subdomain of us."

You're not misreading it: under these terms of service, a link to the front page of the Dallas Morning News Web site is acceptable, but a link to a specific story or material from the archives is expressly prohibited.

Who knows why the site's operators did this. Under any analysis, it's not only bad online etiquette, it's also bad business. Weblogs, search engines, and any number of third-party Web sites drive traffic to newspaper sites by linking to them. Whether that linking is face linking (to the front page) or deep linking (to interior pages) is immaterial.

Whatever dubious goal the newspaper had, the next question is whether the prohibition is even enforceable. There's certainly no law against linking, even against deep linking, but the Dallas Morning News has taken a different tack. It hopes to make deep linking a breach of contract. Under this theory, by reading the Dallas Morning News online, you agreed to abide by the site's terms of service. Those prevent you from linking to a specific article from your own Web site or weblog. As you surf the Web, you stop to read every site's terms of service, don't you?

As recently as a few months ago, you might have read the Dallas Morning News' restriction on deep linking and laughed. But in July, a court in Denmark examined a similar issue and surprised everyone by requiring Newsbooster.com to stop deep linking.

Far From Home

As I mentioned, deep linking is the practice of linking to part of a Web site other than the home page. Newsbooster.com not only made use of deep links, that's pretty much all it did. It was kind of a super-newspaper, containing deep links to virtually all of the major stories in Denmark's commercial press. The Newsbooster.com service was so thorough, in fact, that users no longer needed to go to the home page of any Danish newspaper. All of the headlines from all of the newspapers, with deep links to the corresponding stories, were available in one place: Newsbooster.com. If you didn't want to scan the Newsbooster Web site for articles, the company would even email you the day's list of headlines and links.

And, perhaps most galling of all to the linked newspapers, Newsbooster successfully charged a subscription fee for its service.

The Danish Newspaper Publishers Association, on behalf of nearly twenty member newspapers in Denmark, filed suit earlier this year to stop the Newsbooster linking practice, and it won just a few weeks ago. The court in Denmark ordered Newsbooster to take down all of its links to the sites of the Newspaper Publishers Association's members. A small victory in Denmark, but it still left Newsbooster.com with more than 4,000 other world newspapers.

Web developers could easily dismiss the Denmark decision as irrelevant to the rest of the world, if there weren't some precedent for it in the United States.

Drowning Servers

Online auction aficionados may have fond memories of a service called Bidder's Edge. Like Newsbooster, Bidder's Edge aggregated content from third-party sites into a single place for one-stop shopping. Using automated processes to pull real-time data from eBay and other online auction houses, Bidder's Edge let users monitor and participate in a variety of auctions on the sites of competing auctioneers from a central location. Bidder's Edge presented a gateway into eBay and other auction sites by making extensive use of deep linking.

The mechanism for gathering those deep links, however, was what ultimately killed the service. Because of the volume of auctions that eBay hosts at any given time and the frequency with which the bids for an item change throughout the day, in order for Bidder's Edge to maintain accurate, real-time information on the status of an auction, its software robots had to visit the eBay site pervasively twenty-four hours a day. By eBay's estimates, almost 2 percent of its daily server load went to support the Bidder's Edge robots.

That 2 percent might not have seemed burdensome by itself, but Bidder's Edge wasn't the only online company aggregating auction information. AuctionWatch.com and some two dozen other companies also aggregated links to the sites of online auction houses. In court, eBay argued that if Bidder's Edge were allowed to use robot software to continually mine the eBay site for links, the other auction aggregators would have to do the same thing to stay competitive. The end result would be that either eBay's server costs would have to greatly expand simply to support the commercial efforts of third-parties, or its existing servers would slow to a crawl under the load of these third-party software robots. Given this dilemma, a California court granted eBay's request for an injunction and ordered Bidder's Edge to stop using automated robot software to mine the eBay site for deep links.

All Links Considered

Even venerable National Public Radio (NPR) wasn't immune to the desire to prevent deep linking. Until a few months ago, NPR's terms of service claimed that "linking to or framing of any material on this site without the prior written consent of NPR is prohibited." Forget about deep links, NPR claimed that you couldn't even link to its home page. The only saving grace of the NPR policy was the permission form included on the site. Simply tell NPR your name, email address, home or work address, phone number, name of your site, whether your site is commercial or noncommercial, how long you plan to link to NPR, and provide the words you'll use within the link and on all of the accompanying text on your Web site. After NPR had considered all that, they'd get back to you with a decision.

In June of this year, however, when news of its permission-based linking policy circulated on the Web, NPR came under attack from a Web community that typically was quite sympathetic to the member-supported radio network. An email campaign to the NPR ombudsman coupled with a fair amount of unflattering Internet press convinced NPR to remove its restrictive linking policy.

When questioned about why it had drafted such a draconian linking policy in the first place, its chief complaint was a familiar one. Third-party sites were making such pervasive use of the NPR.org Web site that the source of NPR content was becoming blurred. Because NPR hosted a number of archived audio streams, direct links to the streams from third-party sites would allow a radio archive broadcast to start without the user ever seeing the text, including the credits, that had accompanied the stream on the NPR Web site. (For more information on the legal issues surrounding links to streaming content, see my Web Techniques column, "Linking Legalities," February 2001.)

In other words, NPR faced the same problem as the Danish newspapers. Someone was co-opting its content without giving appropriate credit.

Infringement?

To the best of my knowledge, no court in the U.S. has ever ruled that deep links either infringe on the linked party's rights or are otherwise unlawful. In other words, there is nothing intrinsically wrong with deep linking. As with most things, it's the way you deep link that can lead to trouble, not the mere fact that you do it.

The courts are saying that deep links to specific pages for specific purposes are acceptable. It's a fair use of third-party content to comment on something or link to it. It's also the very foundation of the Web. A court needs a very compelling reason to find that deep linking violates someone's legal rights.

We can see such a compelling reason at work in the Bidder's Edge case. It wasn't the fact that Bidder's Edge created deep links to eBay's site that convinced the Court to shut it down, it was the way in which the links were gathered. eBay had a real, demonstrable problem with server load that only threatened to get worse if Bidder's Edge continued its practice.

Reports from Denmark suggest another rationale for the Netbooster decision. When links into an individual Web site are so pervasive and thorough that they obviate the need to ever visit the linked site's own index and home pages, a court might reasonably conclude that the linking site has co-opted the target site's copyrighted content.

One of the central issues that the courts will look at in the latter scenario is whether the pervasive links create a false impression about the author of the original content. If you've simply created a veneer of links to present someone else's original content, you may have a legal problem.

While the recent case from Denmark was widely touted as the death of the Web, in reality, it's not even close. Few sites are designed to capture a specific third-party site's audience the way Newsbooster did. Most of those creating link tags in HTML shouldn't have to worry about infringement. Even journalists and bloggers, who make links a staple of their online works, should write their tags without fear of ending up a defendant in a lawsuit.

The most troublesome issue though is the contract theory that the Dallas Morning News would use to restrict deep linking. It doesn't purport to base its restrictive policy on theories of copyright, trademark infringement, trespass to chattels, or interference with business relations. It's all about terms of use. You read the site, you agree not to deep link. Click-wrap agreements where a user hits the I Agree button have been increasingly enforced by U.S. courts. Whether they'll also enforce the Dallas Morning News' "click-read" agreements is anyone's guess, but let's hope not. The Web was created for linking. It should stay that way.

---

Bret is an intellectual property and Internet attorney with Hancock, Rothert & Bunshoft. You can reach him at bret@lextext.com.

RELATED ARTICLES Code Signing in Adobe AIR Dr. Dobb's Agile Newsletter 08/08 CUDA 2.0 Released Coverity Intros Software Readiness Manager for Java New Data Center Architecture Proposed		TOP 5 ARTICLES No Top Articles.