A linkbase in hypertext is the notion a separate document or database that stores information about hyperlinks independently from the content documents themselves. Rather than embedding links directly within documents (as in traditional HTML), a linkbase maintains link relationships externally.
In Tinderbox the linkbase is a discrete section of the TBX document's XML inside the <links> tag. Within this each link is stored in a discrete <link> tag. Links are matched either to the ID of a note (Basic links) or for links with text anchors (Text or Web links) to a character offset and range based on a plain text version of $Text (see the <text> vs. <rtfd> tags).
Why use this approach?
Although for most people today, people equate the Web with hypertext, the Web is actually a partial implementation of hypertext due to technical and practical compromises in the early 90s. A full hypertext system would have used a linkbase and Tinderbox follows this model. There is no implicit better/worse comparison with the web, it is simply that the linkbase model is more flexible in a tool that uses a lot of linking, as Tinderbox can.
In addition, a linkbase offers:
- Separation of links and content. Links exist as separate entities, allowing the same content to be linked in different ways without modifying the original documents.
- Bi-directional links. Unlike standard HTML links (which are one-way), linkbases can support links that work in both directions or connect multiple documents simultaneously. Although Tinderbox implements links with a recorded direction (source to destination) they may be queried or traversed in either direction (this is—or was—much harder in a Web-style approach)
- Link metadata. Linkbases can store rich information about links, such as their type, purpose, creation date, or author. For example see the stored XML's <link> tag.
Further background
This concept was prominent in early hypertext research systems like Microcosm and Hyper-G, and influenced the development of XLink (XML Linking Language), which includes linkbase functionality. The vision was to create more flexible, multi-authored hypertext systems where linking could be separated from authoring.
While the web evolved with embedded HTML links as the standard, linkbase concepts influenced various technologies, including XML linking standards and some modern annotation systems.
See also—notes linking to here: