If Les Kelly has his way, image search engines will face tough legal restrictions over what they can list similar to how Napster has been constrained over linking to MP3 files.
Over the past three years, California-based photographer Kelly has doggedly fought against image search engines that have listed his photographs without explicit permission. Kelly lost his most famous battle against Ditto.com back in December 1999, but his appeal will now be heard next month.
Kelly is hopeful that his appeal will be successful, because new judges will be reviewing the evidence and some recent legal decisions have given site owners and artists more protections. In particular, since the ruling in Kelly's case, other courts have found that eBay was trespassed by the crawler from Bidder's Edge and that Napster was infringing on music copyright by linking to unauthorized versions of songs.
"We believe that the Ninth Circuit Court of Appeals will take a fresh look at the original case filed. Judge Taylor [who ruled in the original case” went out of his way to bend 'fair use' so that he would not harm search engines. This feeling was summed up in his statement about 'the established importance of search engines.' We believe that he overlooked the significance of [Ditto's” framing of the full sized image within the results page and the placement of the download button next to the thumbnail. These go directly to harm and have much more significance than Taylor gave to them," Kelly said.
As for Ditto's side, new owners Sorceron, who bought Ditto several months ago, expect that the factors that helped Ditto win the first time will remain persuasive.
"I'm not going to comment on other cases and their relevance to this case," Jason Hardiman, Sorceron's general counsel. "I think the district court was correct in its analysis, and we think that the decision will be upheld on appeal,"
In Kelly's original suit, he alleged that Ditto was infringing his copyright by making copies of his images. Ditto argued this was not true, given that the actual images were not provided to the public, only smaller "thumbnail" images. The court ruled against Kelly, finding Ditto's treatment of Kelly's work to be "fair use."
Image search complaints first came up when AltaVista launched its image search engine back at the end of 1998. It was the first major image search service to present thumbnails, a feature that quickly made the service popular among users looking for pictures. However, this feature also worried artists, who felt AltaVista was encouraging image theft.
Following a wave of complaints, AltaVista added disclaimers to better instruct its users that permission should be sought before using images. More important, AltaVista ceased linking directly to images themselves, forcing users to view images in the context of the entire web page, where copyright notices might reside. It also instituted an opt-out option for concerned artists via the robots.txt file or meta robots tag.
Today, these opt-out options are available from all the major image services that I reviewed below, though some such as Google and Ditto do a much better job of alerting artists to them than others. I also think they'll be key in any court decision, since they provide a way for artists to explicitly say they do not want to be indexed. Some artists, however, may not feel they should have to implement such blocks.
There's also concern that blocking conventions may simply be ignored by some search engines. This is what Kelly says has happened in the case with Ditto:
"[Ditto's” spider did not honor robots.txt. It was a rogue spider that blew past each and every exclusion text that it encountered," Kelly said.
Assuming blocks are obeyed, the mechanisms provide a way for artists to protect their work while balancing the interests of the web community as a whole. If respected, using a robots.txt file is an easy way to block image spiders from AltaVista, Ditto, Google and Picsearch. With some foresight, it is also easy to block FAST's spider while still keeping web pages from a site in the FAST index.
Opting-out via the meta robots tag method is much more difficult, in that only three of the five major services I've reviewed provide a unique command for this tag.
Any concerned artists should take steps to protect their images via blocking in the same way that they might ordinarily institute copyright statements -- and follow up strongly with search engines that appear to be ignoring the exclusion. Sometimes, this may be simply because the blocks were not properly configured.
Of course, whether any type of opt-out ability will ultimately have legal weight may be decided in Kelly's case, which will be heard on September 10, before the US Ninth Circuit Court of Appeals, in Pasadena, California. Even then, the opt-out options may not be tested. They were not a factor in the last case.
Supporting Kelly in his action through friend of the court or "amicus curiae" briefs are groups such as the American Society of Media Photographers, the Graphic Artists Guild, the American Institute of Graphic Arts and the National Writers Union, among others. Behind Ditto stand the major search engines of AltaVista, Google and Yahoo, through their own friend of the court briefs.
Ultimately, it may be financial issues rather than legal ones that restrict image search engines. Ditto's funding is dwindling, while image search engine Diggit, launched last February, closed its doors only months later in July.
In contrast, Google's new image search engine is going strong, but it is essentially a loss-leader for Google. The company says it can afford to maintain it because it is crawling the web already for other reasons, so implementing image search doesn't cost it substantial resources.
Google is also unique in also having a new image search engine that hasn't been targeted by Kelly. Other new launches by Diggit and Picsearch have attracted complaint letters from him demanding $10 million in damages. Why has Google escaped? Because it offers more than image search, Kelly says.
"At this time, my primary interest is in working [against” the pure image search engines rather than the search engines with image searches as an adjunct," Kelly said. "Each has acted, in my own opinion, just like a rustler, attempting to trample the rights of copyright owners, worldwide, in order make a few bucks. They spend a few years creating their own intellectual property and then go out and steal the intellectual property of others in order to make their own intellectual property work for their own financial benefit."
So far, the courts have taken an opposite view -- image search engines are not rustlers but if anything, serving the web as a whole. Whether this view will be maintained gets decided next month.
Want to keep your images out of image search engines? The best route is to use a robots.txt file. This allows you, in one single place, to block several major multimedia spiders.
For example, to keep out spiders from AltaVista, Google and Ditto, you would add these lines to your robots.txt file:
User-agent: vscooter Disallow: /
User-Agent: Googlebot-Image Disallow: /
User-agent: DittoSpyder Disallow: /
User-agent: psbot Disallow: /
These lines tell the image spiders from AltaVista (vscooter), Google (Googlebot-Image), Ditto (DittoSpyder) and Picsearch (psbot) not to index any images from your entire site. The disallow statements can also be modified to reference particular subdirectories or even particular files, as explained more at the Robots Exclusion web site, listed below.
To also block the FAST spider using the robots.txt file, you'll need to place all your images in one location. This is because FAST doesn't have a separate multimedia spider. Consequently, if you were to add FAST as shown for other spiders above, you'd prevent it from indexing anything in your web site, including web pages.
Let's take Search Engine Watch as an example of how to organize and block your images correctly. All of Search Engine Watch's images (more or less) are kept within the /img/ area of the site (http://searchenginewatch.com/img/). This means that if we block access to that particular area, and if all the images are kept entirely within this area, then none of the images will be spidered.
We next need to block access to the area via a robots.txt file with these lines:
User-agent: * Disallow: /img/
That simple two line robots.txt file essentially says "all spiders (the * part), don't spider my images (the /img/ part, the area where images are kept).
As you can see, no particular spiders have to be referenced. Instead, if you organize all your images in one common location, then you can block all spiders at once from accessing that area without naming them individually or without having to worry about your web pages being excluded accidentally.
You can also use the meta robots tag to block multimedia spiders. Unfortunately, Google doesn't recognize this tag for its multimedia spider, while AltaVista and Ditto use different variations.
While Picsearch recognizes the tag, it only recognizes the "noindex" command designed to keep more ordinary search engines from indexing web pages. Therefore, beware! Following the instructions for the meta robots tag on Picsearch's web site will keep out ordinary crawlers that you probably DO want visiting you.
A last problem with the meta robots tag is that you have to install it on each page that contains images, and it will block only the images on that particular page from being spidered. Given this, the robots.txt file alternative is a better option.
If you must use the meta robots tag option, here's what to install:
That tells AltaVista and Google (via noimageindex) and Ditto (via nomediaindex) not to index any images that appear on the page with this tag without causing problems with other crawlers that want only web pages. Support of noimageindex is not mentioned on the Google site, but Google told me directly that it does support the command.
By the way, using spaces after the commas should make no difference. This is definitely the case at Google, it says -- and virtually certain for other search engines. So, feel free to do this, if you prefer:
The meta robots tag can also be used in combination with other commands, such as:
That means not to index the web page (noindex) or images on the web page (noimageindex, nomediaindex).
What if you DO want your images in these search engines? Then don't do anything -- their crawlers should find you naturally. In general, your images will be deemed relevant for particular terms if those terms appear on the page containing the image, in HTML text near the image, in the ALT text for the image or in the file name of the image. Any of these factors may influence the results, while hitting many of them should increase the relevance of your image for a particular term.
The graphical words that appear inside the image itself play no factor. None of the major crawler-based image search engines can read inside of graphical files.
Kelly provides background on his case against Ditto at this web site.
Kelly v. Arriba Soft Corp (Case No. SA CV 99-560 GLT[JW”)
US District Court, Central District, December 1999
Text of the preliminary decision against Kelly in the original case, which was made final in March 2000. Of the four factors determining fair use, the court found two were in Ditto's favor and two were against Ditto. The first factor, "Purpose and Character Of The Use," was deemed most important: "Defendant never held Plaintiff's work out as its own, or even engaged in conduct specifically directed at Plaintiff's work." In other words, because Ditto did not try to pass off Kelly's images as their own or compete in his particular area of photography, Ditto was not deemed as violating fair use, according to the court. It is notable, however, that the court did find that providing an option to see a full-size image on its own violated fair use in the "Amount And Substantiality of the Portion Used" factor. In other words, showing thumbnails was OK, but directly linking to the actual images is more likely to cause trouble for search engines.
Cash-rich Ditto.com returns to home turf
Chicago Sun Times, June 4, 2001
Ditto has closed many of its offices and scaled back in hopes of somehow keeping the company afloat.
Despite its financial woes, Ditto is still operating. Users encounter a copyright disclaimer about image use at the bottom of the Ditto home page and results pages. Selecting an image thumbnail brings up that image in its own window and also opens a new window showing the page that contained the image, where copyright statements might be contained. A prominent "Remove URL" link is at the bottom of the home page
Ditto Robots.txt Instruction
Explains how to remove images from Ditto via the robots.txt file, as well as links to instructions for using the meta robots tag.
Kelly filed a complaint on July 14 with this former image service seeking $10 million in compensation for its use of his images. Later that day, the service closed, with a lack of funds cited as the reason. "The dry venture capital and online advertising climates have finally caught up with Bulldozer Software [owners of Diggit”. Despite Diggit's popularity with end users, we will not be able to continue providing this service," said a notice on the Diggit home page. Undeterred, Kelly is still seeking a settlement or plans to file a lawsuit.
New image search service launched officially earlier this month. As with Diggit, Kelly is seeking $10 million in compensation from the service. A link to a copyright notification page is prominent at the bottom of the home page, and a "You must obtain the appropriate permissions to use any image, linked to from picsearch, from the owner of that material" disclaimer appears at the bottom of results pages. Selecting a thumbnail displays the full-size image in a top frame, along with a disclaimer, and the page it resides on in a bottom frame.
Has a disclaimer at the bottom of the home page but none associated with the actual results. Thumbnails lead to pages that contain the images, not to the actual images themselves. Allows webmasters to exclude images, but a clear link to exclusion instructions is not listed on the home page. Instead, it takes some digging through the help files to find the correct instructions. Kelly initially filed a complaint with the service back when it was unveiled and says that he may still pursue action, pending the outcome of his case against Ditto.
AltaVista Help: Avoiding The Index
Explains how to keep images and web pages out of AltaVista.
No disclaimer on home page or in results, but when you actually select an image to visit, a notice that "This image might be copyrighted" appears below it. The FAQ link from the home page also discusses copyright issues. Images are directly linked to and shown within a top frame, while the bottom frame displays the web page containing the image and any possible copyright statements. Allows webmasters to exclude images, and finding these instructions is relatively easy when reading the FAQ page.
Remove Content from Google's Index
Explains how to keep images and web pages out of Google.
FAST/AllTheWeb.com Multimedia Search
This allows picture searching by selecting the "images" option next to the search box, on the home page. No disclaimers are shown on the home page or results page. Selecting an image shows it in full-size, as part of an image "details" page. That page carries a "This image is copyrighted to its rightful owner(s)" disclaimer. Instructions on saving pictures from results within the site's FAQ area also carry a disclaimer about copyright issues. A direct link to the image is also made available. There is no ability to exclude only images from FAST's database. Instead, if you use the exclusion option via the robots.txt file, you'll exclude images, web pages and any other content on your web site. As with Google, Kelly has not sent a complaint but may do so or take legal action, pending the outcome with Ditto.
FAST Web Crawler FAQs
Information on blocking content from the FAST crawler.
Lycos Multimedia Search
Powered by FAST, this provides access to multimedia files from across the web. No copyright disclaimers appear on the home page or results pages. Following the Help link from the home page does in turn show a link to a "Legal Information" page. This page simply says that Lycos offers no advice as to the fair use of images found from the web. Links lead to the page containing the image, not the image itself. The help link from the home page provides no information about excluding images. This is possible, if you follow the instructions that FAST provides. However, a webmaster would need to understand that FAST provides Lycos Multimedia Search results, and not everyone will make this connection.
Explains how the robots.txt file and meta robots tag work. Be aware that the meta robots tag needs special additional commands not mentioned on this page to block particular multimedia spiders.
How To Block Search Engines
Additional information about using the robots.txt file. It also leads to more information about the meta robots tag.
Multimedia Search Complaints
Round-up page with articles about multimedia search complaints. See the "AltaVista Photo Finder Has Artists Concerned" article listed for more about why artists have unique concerns when it comes to being listed in image search engines and why opting-out via the robots.txt file or a meta robots tag might be a good solution for everyone involved.
Introducing SES Online
Want to view one of the sessions you missed or listen to an especially informative presenter a second time? SES New York sessions are available for purchase on ClickZ Academy's new e-Learning site. SES is now Online!