The Amsterdam hypermedia model

Abstract

On the surface, hypermedia is a simpie and natural extension of muitimedia and hypertext: multimedia provides a richness in data types that faciiitates fiexibiiity in expressing Information, whiie hypertext provides a controi structure that supports an eiegant way of navigating through this data in a content-based manner, unfortunateiy. the concepts that apply to collections of static information do not all translate well to complex collections of dynamic information. What does it mean, for example, to foilow a linic in a hypermedia presentation when the source node consists of nonpersistent data? Does the source presentation freeze, does it go away, do parts of it go away? Similarly, how should the synchronization relationships within and among elements in a composite component be defined? is this part of the hypermedia model or part of a data storage or data presentation modei? in our view, a gênerai hypermedia model needs to be able to specify both the conventionai linicbased navigation elements of hypertext and the compiex timing and presentation relationships found in muitimedia presentations. Such a modei is presented here.Intuitively, support for bypermedia can be defined by taking tbe Dexter Hypertext Reference Model [7,8], cmd augmenting it witb multimedia data types witbin its storage layer. The Dexter model provides a facility for creating links within a document, with the links being anchored inside document components. The notions of links, anchors and components are basic to hypertext systems, and since they are not tied to any particular type of implementation or to any particular type of data, integrating them into a hypermedia system would seem to be a straightforward task. (Note that this does not make the task easy: finding a general way to indicate the presence of a link in a dynamic data, or providing means of "clicking" a portion of video or audio, remain difficult and unsolved problems.) Although such a •Carriage of convenience" can provide immediate results, the underlying control assumptions of the Dexter model make it unsuitable for describing and supporting generalized bypermedia documents.' In particular, this approach cannot adequately support complex temporal relationships among data items, specifications that support high-level presentation semantics, or a notion of "information context" that specifies global behavior when a link is followed-all elements that are of fundamental importance in supporting hypermedia.This article presents the Amsterdam Hypermedia Model (AHM), a general framework tbat can be used to describe the basic constructs and actions that are common to a wide range of hypermedia systems. We present the AHM in the context of the Dexter model's hypertext technology and terminology. We do this because hypertext is-in the time scale of information technology-a relatively mature discipline, with a well-defined model of general system bebavior. The AHM is developed as an extension to the Dexter model in order to capitalize on its contributions to understanding hypertext...

A Novel Approach for Semantics-Enabled Search of Multimedia Documents on the Web

“…As such they do not propose a model for representing rich, structured multimedia documents. In the area of multimedia document models, we find HTML 5 and Flash and abstract document models such as MM4U [18], ZyX [4], and AHM [13]. These models are targeted towards the presentation of multimedia content to the users.…”

Structured Partially Caching Proxies for Mixed Media

“…This structure imposes certain restrictions on the objects, for example saying that two objects should always be retrieved together, after one another, or as an alternative to one another. There are other possibilities as well; a total of 13 were identified in [8]. Knowledge of external structure is important when servers and proxies perform rough synchronization between objects, as irregularities here will affect the user's perception [9] of the media.…”

“…Mixed media has many similarities to hypermedia [8]; both have different media objects and presentation plans. However, the two differ on some key issues: Hypermedia is strongly focused on navigation issues, whereas mixed media focuses more on streaming and concurrent presentation of different media.…”

Multimedia authoring tools: State of the art and research challenges

“…Control flow information is overlaid directly onto the channel view by using a series of synchronization constraint descriptors (called synchronization arcs.) The player allows the runtime view of the document to be presented, and allows user navigation using the link structure of the Amsterdam Hypermedia Model [18]. Note that no explicit programming-based view is used with CMIFed.…”

“…In future systems, the ability to search and 'link' pieces of information will need to become significantly more universal. While the models that support hyper-information are reasonably well understood [16], [18], methods for physically creating links and then activating them need to be improved to non-text media. The associations that are created will need to span collections of related data items and provide support for conditional activation based on the runtime state of a complex document [4], [9].…”