Joi Ito's Web

Joi Ito's conversation with the living web.

Bloggers will already be aware of this, but web publishers are trying to make it illegal to link to pages on their site. The logic from some people is that it subverts the efforts of the publisher to manage the traffic, sell advertising and control the user. National Public Radio say that they just wanted to know who was using their stuff. The great debate following their taking this position seems to have changed their minds. The form one had to fill out in order to link to their page is no longer online and one day after the OJR article, June 27, NPR updated their Terms of Use Page which now says this about linking:

Links to NPR Web Sites
NPR encourages and permits links to content on NPR Web sites. However, NPR is an organization committed to the highest journalistic ethics and standards and to independent, noncommercial journalism, both in fact and appearance. Therefore, the linking should not (a) suggest that NPR promotes or endorses any third party痴 causes, ideas, Web sites, products or services, or (b) use NPR content for inappropriate commercial purposes. We reserve the right to withdraw permission for any link.

Anyway, as co-founder of Infoseek Japan (I'm still on the board), as a newbie blogger and as someone who believes that the contextual flexibility of the Internet is one of its most important attributes... I am horrified by the idea of limiting deep linking. It goes against the basic idea that brought us this medium in the first place.

A court in Copenhagen, Denmark ruled in favor of the Danish Newspaper Publisher's Association against the online news aggregator Newsbooster who was deep linking into the Newspaper's site. This is the first court ruling deep linking to be illegal.

I testified in a case in Japan where an Osaka court ruled that someone who linked to an illegal pornography site was actually running a pornography site and found guilty which has similar legal ramifications, but didn't seem to have the impact that the current deep linking debate has.

Sen proposes an interesting technical solution. (I always like technical solutions more than legal solutions.)

Frankly, I think that if people don't want to allow deep linking, people should just configure their software (probably requires some implementation work and may be some access control depending on how strict the site wants to be) to not permit browsers to reach portions of their site directly w/o a "legal" referer.

It's not a perfect solution, but I bet it would prevent 99% of people from reaching certain parts of websites directly...

Here are some links:

NPR's Mixed Messages - Online Journalism Review
NPR's brutally stupid linking policy - Boing Boing
Public Protests NPR Link Policy - Wired News
Danish Court Rules Deep Linking Illegal - Slashdot
Web site barred from linking to Danish newspaper Web sites -


I like Sen's idea. But could a referrer be spoofed?

Also, Joi, check out this blog:

Some interesting news commentary, and hot scoops.

Since the client decides what referer information to send,
referers can be trivially spoofed (cf. remark about percentage).

I wonder whether people would bother though, and further whether
it's possible to do it via HTML (perhaps w/ Javascript?)

There are several companies actively working on tools for mining dynamic content and deep links. Some are beginning to be pretty successful. Linking is a basic case of the same sort of thing.

No matter what obfuscation mechanism people use, short of crypto, I suspect the problem is essentially intractable.

Perhaps something that looks like a jump table with a computed index might work for some basic cases?

You can tell Bob is into data mining of open sources. I wonder who YOU work for. ;-)

I guess one of the questions in the long run is what is the difference between a database and a web page. Shouldn't we be thinking of more sophisticated content syndication business models than banner ads and traffic analysis within the sites? I want information where and when I want it and I will search out providers that don't force me to conform to their interfaces. This being produced at or marketed at metaphor is not right for everyone, although it may work for some... Does this make sense? I'm breaking my rule of "don't drink and blog"... (I'm having a beer waiting for Mr. David Milstein of Fidelity Ventures who is a bit late...)