5 Trends in Tagging

Gene Smith discussing trends on tagging. He has a book out at http://genesmith.ca called tagging. This led him to identify 5 emerging trends.
Standard model, 3 parts - user, tags, resource. This model comes from Flicker and del.io.us. These are really just the first wave, and they really have not changed very much.
5 Trends
More Structure
Automanual Folksonomies
Leveraging Communities
Rethinking Pace Layers
Sparking Innovation
More Structure
Initially the lack of structure was seen as advantage. This lack of structure allows the difference of the community to become part of the structure, so movie people can be different from the cinema people. Even though this is helpful when it brings more clarity, but structure is wanted when it fails.
Example of structure is wesabe. This is a site where you upload personal data. You have the ability to tag things with tags and sticky tags. Sticky tags will stick to the merchant. So every transaction you have with that merchant will have these tags put on them. The tags can then be used to categorize your expenses.
Talked to the capability to "bubble up" the tags to higher levels, creating a tag hierarchy.
ZigTag - social bookmarking service introduced the concept of semantic tags. As you tag, it references other tags to allow you to provide further meaning beyond "IA". So shows all possible concepts that fit behind the various tags. To create their semantic tags, they mined wikipedia. Making there tags more meaningful.
Leveraging Communities
Example, looking at a tag cloud you can see what a particular community is in to. In some situations the community turns to tags to help define itself. example is LibraryThing. Combining tags gives you the ability to take two tags, and say they are the same. Users now have the ability to create a defined vocabulary. This is not monitored by the community managers. Any paying member can make two tags equivalent, and any paying member can separate tags. Example of Humor, humour - not combined. Used to separate American humor from British Humour. Simple philosophy - only combine tags that are the same in meaning. Interesting concept.
Automanual Folksonomies
Etsy is a good example of this. Place to sell handmade items. How do you design product categories, when you have no idea what people might sell? And each item is hand made. You select an initial top level category, this then creates a set of related tags. These tags are actually the sub level navigation of the site. Library thing is using something similar in a "tag mash" and have used this to approximate the Library of Congress system. They then use this to maintain and suggest tags for new books based on Library of Congress system.
Pacelayering
Concept from Stuart Brand. Idea is that society responds to forces which move at different speeds. From nature to fashion. Peter Moreville applied this idea to IA. Some kinds of IA move slower and are more durable, while others move more quickly. He placed tagging into the fast paced layer. Grant argues that tagging has some aspects of a slower and more durable item. We see that over time people coalesce around a set of tags for content. This is upheld by the paper from the ASIS&T conference from 2006. This is just the power law at work. Example is Buzzillions.com - has taxonomy, faceted nav and user generated tags. They then use the product reviews to create tags and so on.
Sparking Innovation
Architecture of classifying systems. Example is flicker. Any text string as a tag, then you can create a RSS feed based on those tags. Dan Cap is a developer who is a photographer and a coder. He developed a mashup with Google map. He added tags 1) geotagged and two machine tags with lat and long. This let him put photos on the map. In beginning it is fairly primitive - all by hand. Because Flicker architected their tagging system in a nice way, Dan added value to flicker leading to more users, etcetera.
This is happening in many different places - one example is dogear, in IBM, using RSS to tag things into workgroups.
Tags are a fully integrated part of the product and people leverage them - this is what leads to innovation.