Tagging has been one of the coolest ways of managing digital content like pictures, videos etc. It has opened a whole new experience of searching digital content where users simply click a tag (a descriptive word) and every content which is associated with that tag is retrieved and presented to the user.
Tags associated with a content are often provide by the creator of the content. But the true power of tagging is unleashed by allowing users from a community who consume the content to specify relevant tags, essentially crowd-sourcing the task of providing tags. However, this has opened the doors to spammers who specify irrelevant tags to contents. As a consequence, the results of tag based searches are increasingly getting polluted with irrelevant content tagged by spammers. Wouldn’t it be great if there was a simple way to deal with this problem?
How about letting the users of the community to flag such irrelevant tags? A simple “like” button against each tag may just do the trick. Right? Well, that’s exactly the solution proposed by Apple in this recently issued US patent.
But wait! Does it mean that any content provider who allows users to vote on tags would infringe this patent? Although the abstract of this patent may have you believe so, but fortunately it is not the case, as the claims reveal.
The main claim of this patent reads:
1. A method of managing online content, comprising: receiving at a processor a selection of a first item available from an online store; receiving from a member of an online community comprising users of the online store a first tag to be associated with the first item; associating the first tag with the first item; receiving from another member of the online community feedback regarding usefulness of the association between the first tag and the first item; updating an attribute of the first tag to reflect the feedback; and associating a second tag with the first item based on a relationship between the first item associated with the first tag and a second item associated with the second tag, wherein associating the second tag with the first item includes storing the second tag in a metadata record of the first item.
Basically, this claim is on a method having two combined features. The first feature is one of receiving feedback on a tag from a user (e.g. a negative vote) and linking that feedback to the tag. The second feature involves associating tags with a content by borrowing tags from related contents.
For example, assume a picture of Eiffel tower tagged with “Monument” and “Tower” and a picture of Leaning tower of Pisa which is only tagged with “Tower”. According to the second feature, since both these pictures share a common tag, i.e. “Tower”, they share a relationship and based on this relationship, the tag “Monument” may be borrowed and be tagged to the picture of Leaning tower of Pisa.
Interestingly, the two features in the main claim of the patent appear independent and somewhat unrelated. So content providers who only want to clean up tags using feedback from users need not be concerned by this patent.
But anyone who has together implemented the feature of receiving feedback on tags and the feature of borrowing tags from one content to another (which would inherently be based on some relationship between the contents for the borrowing to make sense) would have to watch out.