Disney has been awarded a patent on a "piracy free" search engine technology according to Torrent Freak. The patent was originally filed in April of 2012and describes a mechanism based on "authenticity metrics" relative to a search term that is used to rank results. The presumption is that emphasizing "authenticity" of in presenting search results would diminish opportunities for content unfavorable to the "authoritative" arbiters of "authenticity" to be presented to users of the search engine.
Disney has a tumultuous history of relationships with search engines. Way back in 2000 a Judge ordered Disney to stop using a logo for its online portal go.com due to being incredibly similar to the logo used by goto.com, a search engine. A couple choice bits from the patent application help to illustrate what a Disney search engine might look like in the wild:
For example, a web page accessible from Disney.go.com may be considered authoritative with regard to a search term for tickets to Disneyland™ or a history of the Mickey Mouse™ character relative to a web page for a site that sells tractor parts, but then the tractor parts web page might be more authoritative as to tractors. As a result, one web page may be associated with a high authenticity metric value with respect to one search term, but not another. There may be some sites (e.g., <dishorable-malware-downloader-posing-as-a-freebie-site>.com) that might be of low authoritativeness for all categories and can be associated with a low authenticity metric with respect to the search term (or non-searched terms).
As well as what sorts of things Disney supposes establish authenticity:
In some embodiments, the determination of the authenticity values may be achieved in any suitable manner. For example, authenticity values may be determined based on a set of explicit authentic data, e.g. authenticity metadata, specific web element authenticity characteristics, trademarks, copyrighted material, or web or online sources that authoritative sources curate. Authenticity values may also be determined based on a set of implicit data, e.g. authenticity indicia such as metadata not specific to authenticity, trademark-related information, copyright-related information, sponsorship information, domain information, web element characteristics, references from authoritative sources, references to authoritative sources. Of course, authenticity values may also be determined on both explicit and implicit authenticity data.
While still being incredibly vague over how to determine authenticity in practice:
As an example, authenticity metric values may be determined based on the results of the execution of a fuzzy logic algorithm which associates different weights to multiple authenticity data values based on usage by consumers, third party reviews, and optimization algorithms. As another example, authenticity metric values may be determined using a neural network; illustratively, a neural network may be trained to determine authenticity based on an initial set of web element data and following training, the neural network may be used to automatically assign authenticity weights to web elements. As yet another example, authenticity metric values may be determined by people associating weights to available authenticity data, either transparent to the consumer, or by consumers themselves. As yet another example, authenticity metric values may be determined by regulatory or advisory consortia whom have credentials in the related areas for which authenticity metrics are being derived.
In other words Disney proposes a sort of curated search experience. Filtering out certain kinds of results is explicitly mentioned a bit later in the patent application. Some of the ideas advanced in the Disney application don't seem particularly novel. In particular filtering results based on consultation with advisers is something already done by blekko1 which restricts health related queries to results from an approved list of 76 sites.
Authority and priority in developing search results has historically been a hard problem. Google's original PageRank algorithm was one of the first relatively successful efforts at tackling this problem, but it wouldn't take long for Google's algorithm to be manipulated and the current methods used to derive results for Google Search are far removed from the simplicity of the original PageRank. Over the years Google has worked to diminish the placement of "piracy" websites in its search results,2 but Disney can't be happy with the ordering of the following results that appear above the fold for the search term "Frozen":
Four actual search results are on the page above the fold as seen in this screen capture.
- Frozen (2013 film) – Wikipedia, the free encyclopedia
- Frozen (2013) – IMDb
- Frozen | Official Disney Site
- FROZEN – Let It Go Sing-along | Official Disney HD – YouTube
There's a number of things to dislike about this, but if you particularly dislike that your official web page which serves as a portal to use for promoting and selling a film named "Frozen" and associated schlock is not ranked first. In fact of the 14 links visible in the screen capture exactly one points towards a web domain controlled by Disney. While some people might be frustrated by these results as they occur without any indication that frozen is a word which has had a meaning for some substantial time before the existence of this animated film. Disney though must recognize that in addition to the "anti-piracy" angle often trotted out by media companies, that on the Internet they have almost no control over the presentation of their product.
The patent application is available on Google Patents.
blekko is lead by Richard Skrenta of Elk Cloner, DMOZ, and Topix.com infamy. The people Marc Andreessen will fund never ceases to amaze. ↩
With the notable exception of Youtube. ↩