The paper 'Objektorientert Dokumentdesign i XML - Obdok' approaches aneed for more refined and e cient ways to manage information at thedoorstep of the "Age of Information". It does so by suggesting ananalogy from object-oriented programming to information design. Theparadigm of object-orientation has for almost half a century, provenits favorable quality of giving structure to program code. Informationand information processes lack this quality, and tends to becharacterized by being old-fashioned, one dimensional and linear.These characteristics are traditional characters and reminders fromGuthenbergs book print machine, oral narration style, etc.Information design is mainly in the same situation today as computer science was before objectorientation. Since structural qualities of information will gain more attention in the future, it is natural to try to change information in a way similar how object-orientation once changed system development. Also, when developments lately havebeen released in Dino Karabegs 'Polyscopic Modeling', andtechnologies based on 'Extensible Markup Language' (XML) of W3C thepath is opened for making an analogy between object-orientation anddocument design. The fundamental thesis of this paper describes thispath how to design the object-orientated document language Obdok.
By studying the characteristic properties of object-oriented programlanguages and by finding a way of representing these properties inXML, Obdok has been designed as a language for composing objectdocuments. This model frames the theory of Polyscopic Modeling, and aquantified understanding of information denoted as the 'infon', whichdates back to 'Beyond Truth and Falsehood'**. In Obdok, the concept ofinfon is named by a information unit' that occurs in documents as dierent 'information types'. These information types are represented asXML elements and reflects in a way the 'data types' inobject-orientated programming. Information types, like data types, areclassified into two groups: primitive- and complex informationtypes. Generally, primitive information types cannot containsubordinated element structures, as opposed to the complex ones.
Central in the composing of object documents is the complexinformation type set. This is a heterogenious set that serves both asthe physical and the contextual frame for other information types. Compared to program languages, the set has the role of the object. By applying the di erent information types, this enables theanalogy to be followed closely, also to what superior structuralqualities may concern. Heritage and abstract data types are hencetransformed into heritage and abstract information types. The languagealso specifies a hybrid solution for resource linking. These pointerfunctionalities, vary in the treatments of internal resources andexternal resources. Thus, documents are kept more holistic, preventingspaghetti linking and unstructured information flows.
The fundamental basis for the language is held in a 'Document TypeDefinition' (DTD). This defi- nes the di erent information- andpointer types and decides the legal contents. There is a syntaxspecially developed to declare Obdok documents related to this DTD. Aseparate declaration gives the user the possibility to compose thedocument structure within an object-oriented environment, as analternative to the environment of the writing phase or in a markupcontext. Due to this syntax, it is also possible for the writer toname and structure his chosen information types composing thedocument. This structure is materialized by XML elements in a separateXML file. Declaration and instances are kept separate by initiatingtwo di erent files.
The Obdok declaration and instance files are processed through aPython script to convert and merge these structures into one unified,standardized XML format. This XML structure is made visible throughXSLT transformation to HTML code.
**Jan Egil Hagen, Geir Amdal and Andreas Nergaard, 2000