Home | | Service Oriented Architecture | Precursors of the Semantic Web: Project Xanadu, HyTime

Chapter: XML and Web Services : The Semantic Web : The Semantic Web for Information Owners

Precursors of the Semantic Web: Project Xanadu, HyTime

Bibliography is the art of designing “finding aids” for collections of information. The standard bibliographic model enables a user to find anything, anywhere.

Precursors of the Semantic Web


Bibliography is the art of designing “finding aids” for collections of information. The standard bibliographic model enables a user to find anything, anywhere. Two of the most influential electronic implementations of the bibliographic model are Project Xanadu and the HyTime international standard (ISO 10744). These are the two main precursors of the Semantic Web, which is also an implementation of the bibliographic model.


Project Xanadu


Xanadu would have been the creation of the visionary Ted Nelson, coiner of the term hypertext nonsequential writing. Nonsequential writing is simple: It’s what we do when we throw together chunks of bulleted content in PowerPoint and later put them in the correct order with the appropiate indents.. Of course, Nelson’s vision was a little bit big-ger than that. Imagine, first, that the relationships between bullet points are not expressed implicitly by sequence and indentation but explicitly in the form of an information over-lay that links the bullet points. Second, imagine that each bullet point is on the Internet; and third, that each bullet point has a copyright owner who receives royalties in the form of an instantaneous micropayment every time his bulleted point (or even some characters inside it, according to a patented addressing scheme) is used in someone else’s presenta-tion. Finally, imagine that there’s a worldwide franchise operation, a “Dunkin’ Data” of sorts, that manages copyright issues and micropayments. Now imagine developing such a system before the Internet, before the word processor, before the personal computer, and when the GOTO statement was not yet universally considered harmful. That was Xanadu. “In Xanadu did Kubla Khan, A stately pleasure-dome decree….”


Xanadu was not, it should be noted, vaporware. A lot of labor went into its development; its release date simply slipped, slipped, and slipped into the indefinite future. Xanadu’s legacy has both positive and negative aspects. On the bright side, Xanadu was the con-ceptual pioneer for software that implemented the standard bibliographic model. It set the bar, at least conceptually, for all subsequent hypertext systems, including the Semantic Web. Furthermore, if Xanadu didn’t have all the answers, at least it posed many of the questions that still vex us today: How do we handle rights in an electronic publish-ing environment? How do we manage links, in particular when the endpoints of the link may change or vanish? How do we address into multiple data formats? None of these questions have definitive answers today. It may be a miracle that the Web works without solving any of these posers. However, it is likely that at least one of the questions—How do we handle rights in a global electronic publishing environment?—will have to be answered before the Semantic Web can reach its “full potential.” (Semantics, after all, are a form of intellectual property.)


Xanadu also pointed out a major pitfall that other hypertext systems avoided. Xanadu was conceived as a complete business system; it was to be franchised. Subsequent attempts to implement the standard bibliographic model avoided this pitfall. Both HyTime (discussed later) and Web specifications generally are open and public docu-ments that any business may take advantage of for free. Furthermore, many of the foun-dational technologies of the Web are open sourced. Therefore, Xanadu’s franchise business model seems to be an evolutionary dead end.



HyTime (ISO 10744) implemented the bibliographic model (find anything, anywhere) on the scale suitable for extremely large information owners (such as government agen-cies and manufacturing concerns). Because its addressing model enables the addressing of arbitrary chunks of information and their presentation in arbitrary order, it is also a hypertext system.

HyTime is the oft-unacknowledged intellectual precursor of at least three of the founda-tional technologies of the Semantic Web shown in this chapter and in Chapter 23. It built the foundation for links as information overlays, used graphs as a data model, and enabled semantic links.


First, HyTime was the first markup technology to treat links as an information overlay and to advocate that they be separately stored, rather than embedded within their source or target documents (as in HTML a elements). We see this architectural decision carried through in the W3C XLink, XPointer, and XPath specifications (covered elsewhere in this book). This architecture is also used both by RDF and the XTM effort (XML Topic Maps, covered later in this chapter), both of which specify lightweight information over-lays above sets of resources, rather than being embedded within the resources.


Second, HyTime was the first markup technology to use graphs as a data model. (The graph data model of RDF, and graphs generally, are discussed in Chapter 23.) This was the famous Grove Paradigm, where Grove stands for Graph Representation of Property Values. HyTime faced the problem of addressing into representations of data structures in many formats, not just XML, and without being dependent on any particular program-ming paradigm (such as object orientation). As it turned out, the formal properties of graphs were fit for the purpose of representing most data structures. SGML and HyTime itself were described using Groves, and techniques were developed to describe other data formats. Both RDF and XTM topic maps faced the problem of representing a data model for their interchange syntax (or, in the case of RDF, syntaxes) and adopted a graph for-malism for the same reason.


Third, HyTime was the first markup technology to implement semantic links—links where the semantics of the link endpoints could be more sophisticated and explicit than the simple and implicit “source” and “target” semantics of vanilla HTML. Here again there is a clean line of inheritance to W3C specifications through XLink, but the RDF statement can be perceived as nothing but a semantic link with three endpoints: the sub-ject, the object, and the predicate.


HyTime is sometimes reviled for being too big and too complicated. It is certainly not in the mainstream of Web development. However, where the problems are big and complex, HyTime still finds a ready welcome, particularly in truly humongous technical documen-tation projects for aircraft and/or weapons systems, where its stability, power, and robust-ness really shine. For example, both the European and U.S. standards for Interactive Electronic Technical Manuals (IETMs) are specified in HyTime.

Study Material, Lecturing Notes, Assignment, Reference, Wiki description explanation, brief detail
XML and Web Services : The Semantic Web : The Semantic Web for Information Owners : Precursors of the Semantic Web: Project Xanadu, HyTime |

Privacy Policy, Terms and Conditions, DMCA Policy and Compliant

Copyright © 2018-2023 BrainKart.com; All Rights Reserved. Developed by Therithal info, Chennai.