last change : 30.Oct. 98

French Flag French version 


An authoring environment for multimedia documents

The WYSIWYG paradigm, very useful when editing a traditional document, is not applicable for multimedia documents. Due to the dynamicity of some basic objects and the need to temporally organize them, the authoring phase and the presentation one cannot be merged in a multimedia document editor. This explains why existing authoring environments are closer to programming environments than to documents editors. Nevertheless, it would be a pity that multimedia document authoring be reserved to skillful programmers.

With the MADEUS tool, the goal of the Opéra Project is to design an authoring environment for multimedia documents, which meets the following requirements :

To achieve this goal, we focus on the following directions:

High structuration of the document

A multimedia document is organized according to three dimensions: logical, temporal and spatial. Logical structuration gives the author the ability to organize the document in nested components in order to divide its content in logical parts. It helps him also in focusing on an isolated part of the document during each step of the authoring process (sub-scenes of a multimedia scenario). Temporal and spatial structurations are othogonal dimensions which must take into account the logical one. This way of organizing a multimedia document gives us the possibility to find appropriate specification languages and formatting techniques for both the temporal and the spatial dimensions. Moreover, the logical organization can be used to deduce temporal or spatial information automatically, thus simplifying the author's task. Finally, this is the good approach to allow the definition of generic models for multimedia documents (like SGML for traditional documents). This work makes use of the experience gained by the Opéra project in designing Thot : a structured document editor.

Time Representation

Time representation is required to capture the temporal dimension of multimedia objets like video, audio, etc. and is also used to temporally organize objects with respect to each other. Time representation is clearly the main difference between multimedia authoring environments and traditional editing environments. An appropriate representation of time is the one which, firstly has powerful expressiveness; secondly supports an interface with a declarative approach in the manipulation of documents; finally can be handled easily by incremental and efficient algorithms in order to meet the interactivity of the editing phase and the performance issues required in the presentation phase.

Our choice is based on the temporal constraint networks formalism defined in the artificial intelligence area. It has the advantage to be easily interfaced with a declarative symbolic representation based on the Allen's temporal operators. It is also used as an execution support during the presentation phase. Its structure of graph allows to define efficient algorithms.

Interactive and Declarative Editing Interface

Our goal is to design a user-friendly editing interface which allows, as far as possible, the direct manipulation of the document. To achieve this goal, a declarative approach has been chosen. One of the major problems is to find an intuitive way to represent and to directly manipulate the temporal organization of the document, which is specified by using temporal constraints. We also work on the design of an interactive interface of spatial formatting based on the used of spatial constraints..

People involved in this work:


See the bibliography of the Opera project.

MADEUS is developped in the Operaproject

Send your remarks and suggestions to