Link Structure In The Web

Weblog Kitchen
Welcome Visitors
Things To Do


Recent Changes
Most-linked pages
Whos Who

Hypertext Kitchen

Style Guide
How To Edit
Wiki Sandbox

I'd like to open this area to discussion of link structure in the Web, past, present and future.

Please help me to expand this.


Past: Linking in the WorldWideWeb

Perhaps the success of the Web lies in its simplicity. The HTML standard describes a simple mechanism for connecting Web resources:

A link has two ends -- called anchors -- and a direction. The link starts at the "source" anchor and points to the "destination" anchor, which may be any Web resource (e.g., an image, a video clip, a sound bite, a program, an HTML document, an element within an HTML document, etc.) [1].

Some specific features of link structure in the Web:

  1. Originally, only HTML documents could be link sources. Later, proprietary formats such as Adobe's PDF and Microsoft Word would also provide linking facilities. Still lots of formats (e.g. image, video) which cannot contain links.

  2. Links were embedded in HTML documents, and could not be changed without revising the document.

  3. The TerritorialWeb meant that only the owner could create links leading from a document, and could choose which elements in a document were able to be linked to by other document owners.

  4. With knowledge and server space, anyone could create a new document and link to any other document, without owning the documents linked to.

  5. HTML links had to have a pre-determined source and destination.

Hypertext researchers have often lamented that the Web does not support many of hypertext's rich structuring, navigation and annotation features.

Some common lamentations (PleaseExpandThis):

  • Lack of support for HyperStructures (all links are embedded in Web pages and cannot easily be processed).

  • Lack of support for navigation (only uni-directional, BinaryLinks are supported).

  • DanglingLinks problem (when Web pages are deleted, links which refer to them are no longer valid).

Present: OpenHypertext in the Web

The pervasiveness of the Web has led to the HTML link becoming the de facto definition of the hypertext link. This seems frustrating; HTML offered a limited link structure model despite the fact that exciting advances in OpenHypertextSystems research (see SunsLinkService, MultiCard, Microcosm, DeviseHypermedia, and Chimera) were being made around that time, and 'classic' HypertextSystems such as Intermedia were already well-established - all these systems demonstrated how much richer models of link structure could be achieved.

However, the simplicity of the Web is arguably the factor which enabled it to scale and become the global information repository which it represents today. Far from restricting its success, the fundamental simplicity of the Web has provided a springboard for a multitude of sophisticated tools that add functionality to the Web as a whole, including search engines and server side processing which uses HTML as a delivery mechanism for non-hypertext applications such as online shopping and banking.

Much hypertext research has focused on extending the link structure model of the Web by interfacing a standard Web browser with a powerful hypertext engine. The open nature of the Web infrastructure means that it is possible to extend the hypertext model of the Web to support the more advanced features required of an OpenHypertext LinkService or StructureService.

Some 'hypermedia for the Web' systems (PleaseExpandThis):

See Also: DanglingLinksInTheWeb

Future: Expanding Possibilities

A possible problem with ``hypermedia for the Web'' intiatives is that users must actively seek them out in order to take advantage of the powerful hypertext facilities they have to offer. The widespread adoption of these tools amongst the millions of Web 'weavers' may therefore be limited by ignorance and a lack of understanding of the benefits.

The XLink standard, however, changes the hypertext functionality of the Web at a fundamental level. Advanced link structures are part of the standard itself, so hypertext is a core function which must be universally supported rather than marginalised to third-party applications.


-- Last edited September 2, 2002

Weblog Kitchen | Contact

Sponsored by Eastgate