Presentation

Existing document manipulation systems may be classified into various categories. There are batch formatters [Furuta] and interactive systems [Meyrowitz]. Some formatters such as Scribe [Reid] or Mint [Hibbard] consider the logical structure of the documents they manipulate. Some others, like TEX [Knuth] or Troff [Kernighan], are more concerned with layout, even if macros allow some structure to be introduced in the document.

Formatters have also evolved towards more friendly tools, that allow the user to see quickly on the screen the result of his work: TEX, for example, has several ‘preview’ systems. Janus [Chamberlin] is an original system that has been developed with the same approach. Although they allow the user to see the final form of the document on the screen, these systems cannot be considered as really interactive, as they do not allow the user to interact directly on the final form of the document. Other extensions to formatters have been proposed, by adding a truly interactive editor [André].

Introduction

This paper describes a project within ICL's Office Business Centre that is designing a new document filing and retrieval system. The system is designed to integrate with ICL's networked office product line and to make maximum use of international standards for Open Systems Interconnection. The project is known internally as Textmaster, and an initial subset of the total system is being delivered to selected customers during 1986 under the name ICLFILE.

The system is designed to allow end-users to find the documents they are interested in by means of simple enquiries. They may then view these documents either directly on the screen, or by requesting a printed copy, or by having the document mailed to them electronically. Throughout this process both the typographical layout and the editability of the document are fully preserved. Thus if the user requests a printed copy this can be produced in high quality on a laser printer if required, while if he wishes to edit the document all the necessary layout directives will be preserved. If the document is viewed on the screen it can be presented in a format as close to the printed layout as the screen characteristics will allow: the popular ‘what-you-see-is-what-you-get’ feature of modern word processors.

Introduction

The subject matter of this paper stems from ideas developed in the context of a research contribution made in Lausanne on a document preparation project. The initial goal to produce technical reports has been broadened to solve more general document preparation problems (flexibility, modularity).

As interactive editing systems that include sophisticated typographical features become more fashionable, one might expect traditional formatting techniques to give way. The fact that this is not really the case is due to the advantages and shortcomings inherent in either approach: fast viewing and nice man-machine interface on the WYSIWYG systems, highest typographical quality and greater portability of documents through a variety of textprocessing software on the markup based textfile formatters.

Attempting to combine the good sides of both above mentioned approaches entails several requirements. First, the formatting process needs a flexible parametrisation that provides descriptive formatting specification, clearly separated from the document's content. This approach should offer more flexibility and guarantee portability of a document to several systems with different printing devices. Second, a multi-tasking environment should permit to blend user-comfort with the high typographic quality realized by sophisticated formatting functions.

Introduction

This paper discusses some issues relevant to the design of a Page Description Language (PDL). A PDL is a type of language commonly used for communicating page information from a composition system to an intelligent page printer. These languages are usually specified by the printer manufacturer as an input language, but device-independent outputs from some composition packages, for example DI-TROFF, are also PDLs. I also present a particular PDL, the ACE language, which has been designed by us at Chelgraph and implemented on our Raster Image Processor, the features of ACE itself have been described elsewhere [Chel84, Harris84].

What Is a PDL?

There have always been languages for communicating page information to typesetters and printers. Until recently the capabilities of computer output printers have been very limited, so their input languages have been simple ASCII formats modified by escape codes. Typesetters such as the Autologic APS 5, on the other hand, have quite complex input languages with a syntax and tens of commands.

Introduction

The traditional method of storing information in a printed, linear form as it is done with conventional books has been demonstrated to be inadequate both to represent the complexity of information and to offer quick and flexible access to it [8].

The personal computer appears to be the ideal tool to satisfy these requirements, but a simple computer-based transcription of traditional books is not the best way of taking advantage of the new functionalities offered by computers.

The purpose of this article is to describe one experiment in the design of a documentation system that can take advantage of the flexible data structures and advanced user interface provided by a Smalltalk programming environment.

A semantic network was chosen as the best way of representing the complexity of information ([4],[6]) instead of a more traditional tree structure [5].