»Topic maps are the most powerful metadata standard around.


Understanding Topicmaps



This page tries to explain the sometimes rather obscure concepts used in topic maps in different ways. The goal is to gather different explanations, so people with different mindsets can find at least one that appeals to them.

Please add any explanations you may have...

Explanation for your mom.

So what are these topic maps I'm working on? Well, it's like an index in the back of a book, but bigger. And more complicated. And it all works with computers.

Explanation for InformationArchitects

(This is an explanation I wrote for the SIGIAMailingList) See also TopicmapsAndIA
Topic maps are very cool if you like metadata.

The difference between them and your average metadata system (RDF) is that usually, you'll just assign metadata to an item (a page, ...). But with a topic map you develop the map, and that has a value of its own, apart from any instances it points to. So you may have a topic "commodore 64" and another topic "80's" and you can define a relationship called "was popular in", which has two roles. So basically, you are encoding lots of info in the topic map, even before you have pointed it to your content.

Then it gets even cooler. Say you have a CMS (content management system) and a CRM (client relationships). How can you find out if a certain client, who has bought a certain product (which your CRM knows about) has gotten the latest manual and that new article about the product (which the CMS knows about)?

Simple: you build a topic map on top of the CMS (useful in itself). Then you build a map on top of the CRM (useful in itself as well). Both maps have the same product codes in them. Then this really starts paying off: you now MERGE the two maps. Bingo!

The way to look at them I guess is: they are this really powerful metadata thing (represented in XML) that you can build upon content you have. Then, once you have the map, you can start building applications that USE the map. Topic Maps are (I think) the Future. Imagine this: once you've built a TM, you already have lots of applications available to then start mining that data (since it's an open standard). So you save lots of work there.