Why tag
Tagging — simple classification of resources such as articles, photos or web links — has become an essential part of web navigation.
However, manual tagging is hard. Each time content is added you have to think about what tags to use. You need to remember what tags were used previously so you don't reinvent the same tag under a different name. And if you create a new tag then you need to go back and find all the earlier content that matches so they can be updated. As a result, tags often end up a mess, making it harder for users to follow common topics and to make best use of older content.
Jiglu takes away all this pain. By automatically analysing the content of each page and relating it to other pages it gives you all the benefits of tagging without the hard work. It finds the hidden connections no matter whether they're a month or a year apart.
How does it work
Jiglu Dynamic Discovery typically works with web content or e-mail messages, although it can process any unstructured content. Unlike other automated tagging solutions, it doesn't just analyse pages in isolation: it uses all the content and the existing relationships it contains to intelligently identify what's important.
Content to be tagged is first processed by a natural language engine to extract likely candidate phrases for tagging. A statistical analysis then takes place to determine which should go forward and become real tags, based on factors such as how the phrases are used in individual pieces of content and how they are connected across all the content.
The number of tags to be created is then scaled to the size of the content, so a small set of content gets a small number of tags while larger sets get larger numbers. As further content is added the set of tags is adjusted in real time, with new tags being created when topics take on a new importance or deleted when they are of less interest.
Tags are automatically connected to the items of content they were found in. This uses the importance that was assessed during the tagging process, ensuring priority is given to the items of content that best matche a given tag. Tags can be added manually too, giving the full flexibility of ISO standard Topic Maps but backed up by automated tagging to ensure a taxonomy remains current.
Integration and platform
One of the key strengths of the Jiglu technology is its lightweight integration. This allows it to be deployed tactically where required — it doesn't have to be a site-wide decision. There are three integration options:
- Jiglu Directed Navigation can be added to a site just by adding a single line of JavaScript to its pages. Customisation options allow full branding of the page elements used.
- When greater control over integration and the look and feel is required, there is a simple HTTP API based on the Atom Publishing Protocol. This provides support for automatic and manual tagging of arbitrary content.
- Deeper integration is possible via a Java API, providing support for automatic and manual tagging of content and the full management of a topic-map-based taxonomy.
Built for the Java Enterprise Edition platform, the product supports the Oracle, SQL Server and PostgreSQL databases on the Linux and Windows Server operating systems.
