Many people write to ask me about whether it can be a copyright violation to link to a web page against the explicit wishes of the owner of that web page. This is not a suitable subject for the copyright myths article because this is not at all a decided subject. Like many areas of net law, this question is very much in the air. On Aug 17, 2000, a federal judge in New York (in a decision full of bizarre items) ruled that links were both expressive and functional, and could be subject to regulation. (The foundation I was chairman of was funding the defense of those who did, among other things, the linking, but at this stage we lost, and the appeal was also lost. The case is, however, not over.)

Since I wrote this article in 1996, a number of cases have been tried on this issue. To this point, courts have not enforced terms of service that bar deep linking.

I, and many other people initially had the intuitive feeling that it could never be a violation of copyright to make a link to a web page. Later, I've come to think that this might be wrong in some cases, though only those where you've been ordered not to.

This is not to make a statement about what should be legal or not. Linking is what the web is about after all. A web where you needed to ask permission to link outside of a home page would be a pretty boring one. This is meant more to be an examination and explanation of why the issue is more complex than it seems.

A web server serves up documents to any who ask for them, including copyrighted documents. This is usually set up by the owner of the copyright in the documents, and the owner/publisher wants people to come and pull them out.

The argument goes that if the copyright owner set up the server and delivered the article, how can there be a violation? A link is just a pointer -- and address, and the owner serves the article, not the linker.

The trick is that there may be restrictions on how the pages are fetched that the owner wishes to set. In particular, many web "pages" are really composite documents, consisting of several items, such as graphics and text, and are not meant to be viewed in their individual parts. That they can be viewed as independent parts is an artifact of the HTML language, not the intention of the copyright holders. More particularly, many web pages are supported by advertising, and the owners want people to only view the content if they have also seen the advertising that has been bound to it either directly (by inline into a document) or indirectly, by appearing in a navigation path from the public entry point of a web site.

So does the owner have the right to say you can only fetch pages from a server according to rules they might set? Possibly. Remember that while technologically it is difficult (but not impossible) to stop people from being able to fetch a component graphic from a web page or a sub-page buried under advertising supported menus, the whole purpose of copyright law is to provide legal protection for documents when technological protection is a pain. You don't need legal copyright protection if technological protection is easy, after all.

(That's been shifting, as new law attempts to provide legal protection for the technologies that stop copying instead of just for the works copied.)

So assume for the moment that a site owner has the right to say, "if you want to copy these pages off my server for viewing, go ahead, but only under the following restrictions." Forget about the fact that their server will let you break the restrictions. A book will also "let" you photocopy it, but if the author hasn't given you permission you can't. It's hard to make a book that stops you from doing what the author forbids -- that's why the law is there.

Now the more interesting question -- what about people who make links? How can they be guilty of anything? They just tell people where stuff is. I see two concerns.

The first is the doctrine of "contributory copyright infringement." If you make something whose sole purpose is to facilitate others in violating a copyright, and/or you encourage people to do it, and you know you are doing this, you can be liable for contributory infringement even though you didn't do the copying yourself.

So if you know that the owners of a site don't want you linking inside, and you do it anyway just so people can get around their rules, then I think that if their rules are upheld as valid (unknown) then you could be held liable for contributory infringement. You aren't doing any copying, but you are taking actions solely to cause other people to make allegedly illicit copies for themselves. (This, by the way, is part of what Napster was sued for.) Generally this "deep linking" would need to be costing them money to make a case out of it.

In the DeCSS case, the judge ruled that 2600 magazine, by linking, wasn't providing DeCSS software, but it was "trafficking" in it. What the judge said was that linking wasn't just telling people where something was, which remains legal. Linking, the judge felt, was an action as well as expressing information. He picked the bizarrely extreme analogy of political assassination, which is both expressive and murderous at the same time, to explain that some forms of expression which do other things can be regulated.

The owners of the 2600 web site replaced their links with text URLs, and so far are unpersecuted. This would be in line with the theory that a link is not just a piece of text but also code in a computer language (HTML) that builds a virtual "button" that people push to cause not just speech, but actions to take place.

On the other hand in the TicketMaster vs. Tickets.com, a different district court ruled that TicketMaster could not forbid deep links into its site.

Another way some might think about it is to consider that URLs can be used like access codes. The site owners might give those access codes out to people, but only under their conditions -- ie. if you look at the page with the ad, we'll provide along with the ad the access code to pull out the nice copyrighted info. But does that mean you can hand out that access code to others? You certainly can't if you think of things in terms of a userid/password access code.

But more to the point it is important to realize a link may be viewed as not just like telling people the factual statement about what the access URL for a page is. A link might also be considered as actually the online implementation of a "device" -- a "button" that people push which causes their browser to request a document from another server. You didn't just tell them where the page was, you actually built them a virtual button that, if they push it, causes it to be delivered.

(Of course this is a controversial issue, and one that sits at the heart of many cyber-law issues. Just what is the legal difference between hardware and software, between "real" and "virtual?" We usually think E-mail is mail, digital signature is signature, but many are ambivalent about the mating of the two in some fields.)

What's important about this is that from a legal standpoint, computers don't copy things. People copy things, using computers as their tools. The person who orders a computer to copy something in violation of a copyright is the copyright infringer, and the person who knowingly and deliberately built the system just to make that happen is potentially a contributory infringer.

Confusion of authorship

It's also worth noting that while not directly related to copyrights, other than Berne convention "Moral Rights," the inlining of material or the linking to material on other sites may lead viewers to be confused as to who created certain materials. You might do a site that inlines other people's materials and makes the viewer think they are actually your own. This can be a violation of the author's rights.

It's worth pointing out that it is possible, though a pain, to build a web server so that access to internal pages is done through ever-changing URLs with magic cookies that only appear in the binding or navigating pages. This technique can prevent people from offering links to internal pages or doing inline inclusions of other people's graphics because each URL works only once. You can also check the Referrer field, though not all users -- in the interests of their own privacy -- provide the Referrer field with a web hit.

But should we force people to do this? For one it stops caches from working effectively, which is very wasteful. It also prevents bookmarking by legitimate users. And it's a pain to do when just saying, "hey people, don't do this" seems much simpler and cleaner.

And even if people did make such web servers, you could write a program that went through them, fetching and discarding all pages until the right desired page's magic URL is delivered, and then used that to avoid the other pages. Should this be legitimate?

Note that ad-blockers are not necessarily illegal as long as they have a legitimate purpose. So while something that deliberately removes the ads from Yahoo might be illegal as contributory infringement, a general program that removes "all top of web page GIFs 460 by 55 in size" would not be.

Inlining

The quesiton of inlining, may, as noted, be treated differently from linking. Inlining involves writing HTML to insert an image that is fetched from another server. Thus when people view your web page, they see the image from the other site and the other site actually does the work of serving it.

Inlining is strange because in some ways the law might treat it as a lesser offence than copying the image from the other site onto your server and using it there. In that case, you would be engaged in a very straightforward copyright violation. With inlining, the original image owner's site does the work of copying the image to the user viewing it.

However, that seems backward, since the effect is the same, and to add injury to insult, not only are you seeming to use their image without permission, you're making them do the work. They're getting the bill for web service -- and images are the bulk of web traffic at most servers.

The doctrine of contributory infringement might apply more strongly here because you've now set up an automatic mechanism to cause users to copy the picture in a way not approved by the copyright holder. They don't even know it's happening, it's so automatic. It might be akin to you placing a photocopier on the street with a copyrighted photo in it, and a sign saying "press this button for a free picture." And sending the bill for the copying to the copyright holder of the picture to boot. You don't make any copies yourself, but it certainly seems to be even worse, from the copyright holder's viewpoint, than if you had done so.

Framing

Another interesting issue is the concept of framing. That means building a multi-pane web page where another person's web page is contained inside a frame with your own content -- such as your own ad. You see a page from the New York Times web site, but somebody else's ad -- an ad sold by the people who pointed you at the NYT web page, not the NYT -- sits on top of it. There are lawsuits going on over such activity right now.

One interesting issue about framing is that it is similar in some ways to the activity of advertising supported ISPs. At an advertising supported ISP, you get free internet access, but you have to use a browser they provide which runs ads on your screen while you are logged on. So in this case you might also see a NYT web page with an ad sold by the ISP sitting on top of it.

I think the ISP should be legal but the framer should not. The copyright infringement here is the creation of a composite or derivative work without the permission of the author of the main component. The question, however, is who created that composite work? Who joined together the ad and the web page? In the case of the framer, it was the framer who made the connection. In the case of the ad-supported ISP it was the ordinary browsing user.

Permission

Note that above I'm talking about linking against the express wishes of the copyright owner. Ie. the owner put up a notice saying, "Don't link here" or asked you not to link. Without such a notice, there is a fair argument that the authors of web pages welcome links to them and give implicit permission for this to happen. The law will eventually rule consistently on that issue, and if it decides against the ability to link, it will no doubt revolve around what a reasonable person would assume about the intentions of the web site holder.

But my guess would be that if you have an ordinary web site, and linking is not going to bypass some security, or payment system such as advertising, and there's no information anywhere about the site not wanting you to link in and no reason to believe they don't want it, linking should be very safe.

Certainly you can feel free to link to these pages!

There are other items on my pages on copyright, including:

You can also read my radical new philosophy of intellectual property