Comments on: Introducing Supertags: Advanced Web Taxonomy

By: Ardekantur / A Basic Supertag Implementation

Ardekantur / A Basic Supertag Implementation — Sun, 25 May 2008 04:56:45 +0000

[...] last time I talked about Supertags, but now we’re actually going to see an implementation. This is [...]

By: Jamie

Jamie — Fri, 16 May 2008 16:13:55 +0000

To me this seems like a narrow subset of concepts related to work being done on the semantic web. An interesting topic with a lot of potential for sure.

Have you considered an automatic system of tagging based on the language of article/item? To me this seems like the most exciting aspect, once the algorithms can be sufficiently tweaked to accurately reflect the contents meaning. Starting with simple tag suggestions by looking for common relationship building clauses.

The more we can do to help machines under the underlying meaning of our human generated content, the better. Keep up the good work.

By: Ardekantur

Ardekantur — Fri, 16 May 2008 06:01:57 +0000

I am defining the relationships between tags, but what I envision is that those relationships, and the tags on either side, comprise one giant chunk of information that can be applied to an item. Some relationships may also implicitly say that both terms would be applied to the item as regular tags: if an article is about Ruby being better than Python, then the supertag ‘ruby>python’ would not only define that relationship, but apply the tags ‘ruby’ and ‘python’ to the item.

By: Robert 'Groby' Blum

Robert 'Groby' Blum — Thu, 15 May 2008 22:58:35 +0000

I’m not sure if “tags” is the right name for this. You’re defining relationships between tags.

The idea of attaching these relationships to individual articles, however, is a pretty cool one. In a way, it’s suggesting a formalized way to summarize articles.

The big problem I see is that you have a limited vocabulary to express those relationships. (How many special characters are there, after all). As a result, you will lose the intuitiveness that tags have.