Tagging

Playing around with web2.0 ideas on a side project, I'm running up against the "strong-typed" data vs. "tagging." This is similar to the taxonomy vs. so-dubbed "folksonomy" thoughts.

Folksonomy Basically, a Folksonomy, as its name suggests, it's like a taxonomy, but instead of hierarchical, it's generally (in web terms) created by a community through "tagging." So you end up with a less structured result, but you may end up discovering things "from the wisdom of crowds."

For example, if I applied a taxonomy-style categorization of movies to say, 2001: a space odyssey I might come up with things like:

• Sci-Fi (genre taxonomy)
• Kubrick (screenplay / director)
• Oscar Winners (class)

Tagging would probably get similar tags to the above categories, but you might also get things like:

• Classic Sci-Fi
• HAL 9000
• Tang
• Monoliths
• Artificial Intelligence
• Meandering 60s movies

Back at the Ranch I haven't decided how to implement one or the other. Should I give my users strongly-typed fields for them to enter? Would a user prefer to have a field for tags and obliterate the other fields? If I give both, will it get confusing as some users use fields and other no fields and tags - how will they be able to find data?

These are important questions to answer before designing programs, websites, or even forms, questionnaires, and the like. How are people conceiving data? The better you match the input options to what is in their brains already, the more quality information you get from them.

At least partly due to this internal debate, the Google Video: " Nurturing Tagging Communities" caught my eye. In it, he talks about different "classes" of tags (factual, personal, subjective) and also how displaying existing tags to user generates more of each type. Graphically (@20:25):

His disclaimer: "We can't be sure that these results are deterministic."

This adds some relevant information to the questions about folksonomy/taxonomy. Namely:

1. People are ready to add a lot of useful-to-others (factual) tags
2. You can set up your application to get more factual tags (thus lowering the cost of not using taxonomies) premise: if you have good solid tags, you aren't losing much by not having a taxonomy (like the 2001 example above)
3. People like tagging - particularly they enjoy applying personal tags. Ok, this isn't from the graph, he says it elsewhere in the video. Interesting: there is a user preference. If you app is about self-expression, personal tags are seen by users as part of that expression.

Final Most of my thought around this is very centric to the work I am doing, so it is somewhat technical (in a software design way), but I think the question of folksonomy/taxonomy is interesting - I think we will see more of this as time goes on. Similar to how we saw TV ads with more vector-based animation - pop culture imitating the web (early Flash work had a very particular style), we will likely see more "open-ended" classifications of things in "the real world" as people bring the sensibilities and thought about "tagging" into other arenas.

I don't doubt this is a shift in peoples' minds that is happening. I'm just wondering how to deal with it and how to best utilize it.