Third-generation distributed hypermedia systems are not yet available!

WWW

• Most widely used distributed hypermedia systems, but the WWW lacks features that have been part of earlier second-generation hypertext systems.
• Developed by Tim Berbers-Lee at CERN in the 1989-1991 period as an information system for collaborating physicists.
• Transcends hardware and operating systems dependencies
• Hyper Text Transfer Protocol (http) is a simplified file transfer protocol with extensions for hypertext operations.
• Required documents are transferred in its entirely to client's system for viewing.
• Non-text media such as image files, audio, and video may be referred in an HTML file, which are retrieved in bulk after the client has downloaded the HTML files.
• Format negotiation algorithm
• Clients can specify several acceptable alternative formats for a given object in addtion to maximum transfer time and a maximum object size.
• Server selects the best format given these constraints

Figure: Client and server components of the WWW architecture

• Java introduced a number of important ideas for extending the WWW
• A model for secure application deliverly
• Incremental deliverly of applets
• Platform-independent delivery of applications

• Distributed hypermedia system developed at the Institute for Computer-Supported New Media at the Graz University of Technology
• The designers describe it as a second-generation distributed hypermedia systems
• Each Hyper-G server maintains a hierarchical organization of its documents. This hierarchy, based on a hyperdocument node type called a collection.
• Each Hyper-G server has a document server, link server, and full text server
• Document server provides Hyper-G documents to the presentation clients
• Link server stores and processes links separately from their documents, facilitating the use of link overlays and enabling the storage of links in a database- > the documents processed by Hyper-G can show the links that lead to them as well as from them
• Full text server is a text search engine that permits full-text search of Hyper-G collections both within a single server and across server boundaries.
• Hyper-G uses a client-server protocol similar to http
• Clients connect to only one home server
• Home server acts as the gateway to all remote repositories including non-Hyper-G servers, such as gopher, WWW, and wais.
• Browser: Harmony viewer

Figure: Hyper-G server architecture

• Open and extensible hypermedia architecture developed at the University of Southampton
• Allows users to browse and query large multimedia collections
• Stores links separately from documents
• Has ability to store and process links connecting to information provided by third-party applications as well as connection to information stored in documents
• Orchestrates multiple document viewers for a single presentation
• Coordinates multiple filters, each of which is a source for information and external processing that makes up part of document
• Controls many application that can act as viewers through Universal Viewer (UV)

Figure: Microcosm distributed model

• Hypermedia engine for HyTime
• Developed by the authors of this chapter
• The document model is open to any HyTime or SGML DTD, including HTML
• Hypermedia semantics of HyTime related to links, time and space, and anchoring are available
• The server stores a parsed compact version of document, simplifying the design of the client, allowing incremental document access, and leading to reduced network overhead
• Design approach comparing with the current httpd server:
• Saves significant storage space at the server
• Eliminates the need for client side parsing
• Reduce the amount of information being transmitted over the network
• Is the basis for incremental delivery of SGML-encoded documents
• Facilitates server-side document structure queries