Every now and then I come across a blog that has really great information. Tagsonomy is one of those blogs and in particular I am really impressed by this post: The year in tags Having all these items together in one post really reinforces the changes that are happening at the moment.
IMHO what is missing is the ability for tags to be compared and matched, so if something is tagged in Amazon, Yahoo and Del.ico.us then searching for a tag on wordpress.com reveals content on those other systems – because the various tags are known to be the equivalent of each other.
I'm not sure if that was clear enough. You have probably seen opmlsearch.com right? You can search for a keyword there and find the opml files that contain that keyword. Cool. Is there such a thing as a tag search?
Update (Tues 4 April 2006): I found a tag search site! KeoTag
If I go to a tag search site and look up a tag, will it only return the posts that have that exact tag?What about related tags? What about similar tags? If you're only finding that exact tag, how is that different to a keyword?
Tags are not keywords. Content can be tagged differently on different sites. How we choose to tag content can reveal much about the content and also about our own thoughts.
Tags help to reveal two things:
- It reveals my perception of the context of the post I am reading. It's about the post. "I think this is about search" or "It's about memes" or "It's definitely about tags", or "What a great post about keywords"). A tag is in the eye of the beholder. (Or is that beauty?!)
- It reveals my perception of the thing *I* am looking for. It's about me. I am looking for "food","meat", "a recipe", "research" or whatever.
It's the same piece of content, but my choice of tags can reveal my intention (about me) and can reveal my perception of what I am looking at (the object).
So a tag search site would not function like a keyword search engine (simply looking for a keyword), it will have a deep understanding of core root or stem words and ontology in order to understand what the implied context is (of the post) from the reader's perspective.
When the tag search system can correctly understand the various tags that have been applied, then it can far better understand the post.
So here's the parts of what is needed for a tag search site:
- We need to have access to all tags on all systems (APIs from Amazon, del.ico.us, technorati, google etc.)
- We know the intention/perception of the person who tagged the content (relatively easy!)
- We then need to understand the intention/perception of the searcher (hmmm?!)
- Match the two together for a perfect results set containing content which matches my intention/perception
This would result in a far better search engine than what we have today.