Configuration Control for Product Documentation

A Way of Integrating STEP & SGML

Page 7

Annex 2

SGML Representations & SGML_String

Overview

The preceding sections have discussed the

For an overall integration there is a need to establish information models for the following domains (beyond product definition models):

1. the representation of an information object by a presentation element structure, e.g., paragraphs, lists etc., which provides the syntax for the correct mark-up of the content of an information object. The presentation element structure is referred to as the sgml_representation, whereas the items forming the representation are referred to as sgml_representation_items. For example, an sgml_representation may be a DTD or a subset of the DTD, whereas an sgml_representation_item may be at the level of an element declaration within a DTD.⁽¹¹⁾
2. in order to support the organisation of text within information types, element structures may be used to define the mark-up that may be used within an instantiated information object. Such element structures do not convey any meaning beyond the use for the organisation of text in paragraphs, items, lists, tables, etc. It is the intention of this document to propose a unifying model for such structural elements for common use across all DTDs to enable re-use and sharing of information objects across various publishing structures.⁽¹²⁾
3. the name of an information type represented as an SGML element structure must reflect the underlying meaning of the information being represented. This will be an alternative to the underlying meaning of the counterpart in the EXPRESS entity definition. An information model is required to represent such underlying concepts. It is recommended to write such an information model using the EXPRESS information modelling language.
4. When information objects are published within a framework (such as a document), the structure of such a framework may be a separate information model. The model can be implemented in SGML or in a database environment.

Planning Model for Configuration Control

The subject area of Configuration Controlled Authoring and Publishing may have the following schema level planning model:

Schema level Planning Model for Configuration Control Authoring & Publishing

The presentation element structure representation schema contains the element structures (SGML markup definitions) to encode the content of an information object. (Note: This is an equivalent approach to STEP part 42 which provides the standardised syntax for geometric definition elements for a spatial representation of a product.)

The information type representation schema contains the data types, which represent the classes of information content of product documentation. The set of common information types as identified above will be defined within this schema.

The framework representation schema contains the definition needed to configure information objects for publications. For example, it may contain a model for traditional documents defining concepts such chapters, footnotes, index, etc. just as it may contain models for job cards, electronic messages, etc.

The Configuration Control schema contains the facilities to manage versions, approval, change orders, etc. for information objects and publishing structures. In traditional document management, this information domain would contain the definition for document management.

In context of this document, the product definition schema is a placeholder for any relevant STEP model, which defines product structures, properties or other information concepts relevant in the context of current STEP APs.

Data Planning Models

Information TYPE Representation

The principle content of this information model is focused on formally establishing the concept of an information object and to classify information objects with a standard classification schema, as an information types.

Any information object should only be classified once.⁽¹³⁾

Fig. 9 illustrates the frame for the Information Object Representation using an EXPRESS-G partial entity level model annotation:

Information Object Representation

The concept of an information object is established as an EXPRESS entity type. For an unambiguous definition of the keyword, a modelling framework is used. A modelling framework was defined in the Process Industry for the development of STEP Application Protocols for that industry sector [EPISTLE]. The EPISTLE framework model is applied to interpret the term information object by stating the following about an information object:

• · represents a perceptual structure of information content to convey meaning for humans
• · is a model of a specific thing, an occurrence
• · is an actual thing;
• · is a fact

The information object can be classified (is type of) an information type. The concept of an information type can be specialised in orthogonal concepts as outlined in a previous section. In fig. 9, the shown inheritance structure for information type represents a partial view regarding the possible classes of information type.

The content of an information object is represented by a complex data type, referred to as sgml_string. A model of sgml_string was published in an earlier T14 White Paper [ref.????} and is shown below.

SGML string

In an earlier White Paper [SWEDCALS] the concept of SGML_String was introduced for representation of textual information in context of STEP based product model. The basic model of SGML_String is shown in the figure below. SGML_String models marked-up text. The structure of the SGML based definition is not part of the SGML_String concept, but references are specified for the proper use of such definition, for example to validate the content being provided in an instance of SGML_String.

The relationship between and instance of SGML_String and its classifying Information Type is included in the model of Information Object.

Figure 10: The SGML_STRING

Presentation ELement Structure Representation

The principle content of this information model is to define standard SGML element definitions for re-use in various Document Type Definitions. For example, an element structure can be defined as a paragraph definition. For completeness, this information model contains:

• ·a model of a subset of the SGML language which addresses the declaration of SGML elements, attributes, and entities.
• ·a model of a document type definition being built up from declarative statements and an implicit or explicit defined SGML declaration.
• ·a model to classify SGML elements or groups of SGML elements in accordance with the semantic content of such elements.

The below diagram represents a partial entity model in EXPRESS-G in order to illustrate the domain of information.

Figure 10. Presentation element structure representation.

The concept of an element type declaration, entity declaration or attribute declaration is represented by a common abstract supertype dtd_declaration_item, of which those declarations are specialisations.

⁽¹⁵⁾The above diagram provides hooks to specify start tag and end tag minimisation without giving any specific details.

This model is only developed to a level of detail needed to show the syntax of an SGML element type declaration. The content model of an SGML element type declaration is modelled using an association indicating that the content model of an element type declaration can contain other (sub-)elements or members of the sgml_base_type. The attribute to the association entity indicates the sequence of sub-elements within the content model as well as their usage flags.

The entity type sgml_base_type is a hook for representation of more specialised concepts such as PCDATA, CDATA, EMPTY, etc. without providing any details.

To ensure reuseable DTDs or portions of those, an SGML element or a group of SGML elements can be classified by giving such an element or a group of elements a "standard" name. This means, that the name of an SGML element or the 'root' of a group of SGML elements is the name associated with its class. These names will follow common constructs such as 'list', 'table', 'paragraph', etc. The class concept is referred to an declaration form.

The model above allows the representation of an SGML element type declaration or a group of element type declarations to be managed as instances of the above model or as SGML defined (partial) DTDs outside of the context of this model.

Framework Representation

The principle content of the framework representation schema will be the definitions for publishing structures, version management of such structures etc. If we would name a ready-to-use publication, such as a maintenance or a repair manual, as an information product, this framework representation schema would contain the equivalent information product definitions as Part 41 of STEP provides for the definitions of physical products and their components.

Fig.11 provides a high level overview regarding the content of the framework representation schema:

Figure 12. Framework representation

The core of this data planning model is the concept of a structure element, which models an instance (occurrence) of a structural aspect of a publication as e.g. a front matter, body, abstract, footnote etc. The content of a structure element is an information object, such as the content of an abstract, a footnote etc. The information object is already defined in the information object representation schema above.

Structure Elements contain other structure elements or information objects. The entity type composition of structure element is an association indicating that an information object or another structure element can be part of. The attribute sequence of this association ensures the proper order of the parts within the whole.

An structure element has a DTD declaration associated with it. The entity type component DTD provides the proper declaration for instantiation of an structure element. The definition of component DTD is already given as part of the basic markup representation schema.

Product Definition

In context of this document the product definition schema represents a placeholder for any product definition structure of STEP. As illustrated in the below model, an information object may be authored for functions, relationships between functions, products, relationships between products etc.

Figure 13. Product Definition & Informaiton Objects

Configuration Control

Again, in context of this document, this domain of information modelling is a placeholder to describe release, approval, and change order management. This document assumes, that existing concepts from the development of STEP information models can be re-used.

Extensions into other Representation Forms

Another form of management of user data is in the use of database management systems. An implementation of user data in a database management system may be specified by use of SQL. Although relationally oriented, SQL is another product independent way of specifying structures and data by use of a standardised language.

Because the implementation of STEP for data sharing becomes more important over time, the mapping between an information model in EXPRESS and its mapping into a logical model of the database in SQL will also be addressed in this paper.

Furthermore, database technology becomes increasingly important to store SGML information in a persistent form for data sharing. Document type definitions must be mappable from a logical model of a database, as well as data translation & transformations must be performed between the SGML instance data and the implemented database environment. It seems as if the database environment is becoming a common form or implementation for STEP-based and SGML-based information, in which STEP as well as SGML provide exchange mechanisms for such database environments.

The above comparison between SGML and EXPRESS is extended to include SQL-based logical models and data translations & transformations across such environments.

Footnotes:

⁽¹¹⁾ This needs some discussion. The item level is correct but the sgml_representation may be better reflected as a publishing structure.

⁽¹²⁾ this is now documented in the preceding para.!

⁽¹³⁾ I'm not certain that this statement is either correct or relevant.

⁽¹⁴⁾ There needs to be an element name (Generic Identifier) associated with the element type decl. and the content model, i.e., the decl. associates an GI with the model.

⁽¹⁵⁾ I'm not certain that this statement is either correct or relevant.

Previous Page | Next Page | Table of Contents