Thursday, February 23, 2006

Google denuded

The soon to be much talked about Google Images opinion is available here. Under this decision, the court granted Perfect 10's request for a preliminary injunction. The court found Google's use and creation of thumbnails an infringement. However, I'm not sure that that's the most interesting part. Since I'm a former web developer, let me look at only the deep linking/framing issue.

The plaintiff, Perfect 10 is a proprieter of "high-quality, nude photographs of 'natural' models." (They seem to have a thriving business selling small images for cellphones in the UK--about 6000/mo, interesting...).

Google's Image search engine creates thumbnails of the images which it presents to users searching its database. When Google returns a result (like this for Julie Bowen) clicking on a link takes you to a Framed version of the website. The frame, on the top half shows the cached thumbnail, and below the page from which the picture originated.

The court analogizes "in-line linking" and "framing". First, the image on the right of this paragraph of William Patry is "inline linked". That is, in my HTML code I have told your browser to fetch that image. It's not stored on my server--well it's actually on blogger.com, but that's not really the point--and is not actually part of my website. This means nothing is stored here, and nothing is sent to your browser except the information disclosing the location of the image. Second, the box immediately below this paragraph is a "frame" of William Patry's discussion of this Google case. His page is wholly contained within my own:

In determining whether such activity was direct infringement, the court applied the "server test". The test merely asks, "on what web server is the content actually stored?" Under this test, neither of the above examples would constitute direct infringement. Similarly Google could directly infringe simply by linking to an image (including directly linking to full size image and showing it on its servers) or by framing Perfect 10's content.

Since Google merely frames full-size pictures located on Perfect 10's on site, Google is not directly infringing those copyrights. However, Google loses under the "server test" as applied to the display of thumbnails on its search result page.

The problem

As a former web developer, I think that that the decision is largely correct. ON the facts presented to the court, it clearly must be correct. However, nearer the margins I'm not so sure. Let's consider a few things first.

The test for direct infringement is really a question of "where the content 'originates'". Since neither the image of Professor Patry nor the content contained within the frame reside on my server, I'm not infringing. To the unknowing, the fact that the image isn't on my server would probably go unnoticed. Can it be that the test for infringement lies merely what's in the source code? Is it only HTML source code?

A number of web development languages allow developers to "get" content from another website. What if a developer "fetched" the infringing content and merely returned it to the browser. Nothing stored (at least for very long) on the server but everything returned directly with the page as it's sent to the browser. Many small sites have used a similar techniques to add weather/stock tickers to their pages. If one were to look, the only thing found on the server is the "request" to get the information from those pages.

That sounds like direct infringement to me. It may be that under the "server test" it would be as well. Since the content makes a momentary stop by the "infringers" servers, you've got infringement. However, how many other servers would that same content momentarily pass through on its way to the end user/browser? A trace route of any website from your computer reveals that there are probably dozens of computers between you and the content.

What's more, I think that Google gets a break because of the obviousness of the fact that it did create the content. The search results in fact reveal the actual source. That is not clearly the case of "inline-linking".

0 Comments:

Post a Comment

<< Home