XML in Investment Banking		Table of contents		Indexes		A Healthy Revolution

Using AECMA 1000D/ATA 2100 data-sets to generate Class IV IETM's

Ronald Walraven

Consultant

Stork TPS Industrieplein 3
7553 LL Hengelo The Netherlands
Phone: +31 74 240 3700
Fax: +31 74 240 3777
Email: walraven@tps.stork.nl Web: www.tps.stork.nl

Biographical notice:

Eurocopter

France

Marignane Cedex
Raust, Bernard

Ronald Walraven a is Senior Consultant working for Stork TPS. He is and has been project leader of several IETM implementations, including the discussed IETM for Eurocopter. Current activities cover SGML and XML implementations at different customer sites. His main interest areas are optimising service and maintenance activities by means of information, techniques like SGML, XML and the different style sheet languages and electronic distribution of technical information.

Bernard Raust

Senior Engineer

Eurocopter Etablissement de Marignane
Marignane Cedex France 13726
Phone: +33 44 28559889
Email: bernard.raust@eurocopter.fr Web: www.eurocopter.fr

Biographical notice:

Bernard Raust is a Senior Engineer at Eurocopter, the Franco-German helicopter manufacturer, where he is project leader of INDOC an IETM product able to integrate ATA2100 and AECMA1000D Technical Documentation. He is member of Electronic Publication Working Group which is in charge of defining IETM policy for AECMA (Association Europeenne Constructeur de Materiel Aeronautique). His activity interests include STEP modelisation for integrated logistic support information systems, knowledge based systems and artificial intelligence, user interface and advanced communication and media systems.

ABSTRACT:

In 1998 Stork TPS and EuroCopter, the Franco-German helicopter consortium, formed a partnership to develop preparation tools to automate the production of class IV IETM’s using ATA 2100 and AECMA 1000D data. The presentation and this paper contains details of the experience and knowledge built up during the development of these tools.

Introduction

In 1998 Stork TPS and Eurocopter, the Franco-German helicopter consortium, have formed a partnership to develop preparation tools to automate the production of class IV IETM’s based on either ATA 2100 or AECMA 1000D data. The presentation and this paper contain details of the experience and knowledge built up during the development of these tools. Attention will be paid to the downgrading (or is it upgrading) of SGML data to XML.

We start with a short introduction to the concept of IETM’s. Then we elaborate on the data preparation of AECMA 1000D and ATA 2100 data to satisfy the specific requirements for an IETM production. We take a deeper look at the specifics of the two types of data and discuss the specific IETM requirements before we discuss the actual data preparation. The paper ends with a discussion of the results and some conclusions and remarks.

This paper uses the term "he" for a person. Everywhere "he" is used "she" can be read as well.

What is an IETM and what should it do?

To simplify the explanation of the concept of an IETM the American Department of Defence (DoD) has introduced a classification for IETM’s.

Classes 1 to 3 basically are still an electronic representation of the linear manual concept. Besides the electronic availability, supporting search mechanisms and internal hyperlinks, these types of manuals do not offer additional intelligence to the data.

Class 4 however moves away from the linear approach and gives a base for a non-hierarchical approach to organise the information. Information has to be broken down to smaller grains of information, the so-called data modules. Next to the organisation of information the requirements for the browser are more extensive as well.

The DoD definition, however, does not state what type of interactivity the IETM should provide.

An IETM as an electronic form of distributing information has major advantages over traditional paper based publications.

For the user these include effective navigation through the data set, multiple ways of retrieving information, such as full text and index search, and the use of hyperlinks to facilitate instantaneous access to related information of any form. Also the inclusion of non-textual data such as video, audio and (colour) graphics offers great possibilities to further increase the effectiveness of the information. The information can be presented in such a way that the user only sees the information required for the specific tasks and equipment he has to work on, dependent on his experience level. The user can be supported with intelligent fault localisation and with functionality supporting the logistics around the service process. All of these features of an IETM help the user to perform the service tasks more efficiently.

Last but not least an IETM enhances the image of the company which may be of extra importance when one has to keep up with the competition.

There are some obvious reasons for the distributor of the information to create an IETM. A physical CD-ROM is inexpensive to produce and easy to distribute to the end-user.

Other reasons are less obvious. A good IETM improves the efficiency of service. The data can be accessed from multiple points of view. The data can be scaled to users and tasks.

With hyperlinks the user can easy jump to related information, without a second of searching.

IETM preparation for AECMA 1000D and ATA 2100

In the sections below we discuss some of the most important issues of the development of preparation tools for an IETM class IV based on AECMA or ATA data. The goal was a tool that can take a complete set of data maintained by in this case Eurocopter and converts it into an IETM with specific requirements. First the specifics of the data and the specific requirements as to the IETM are discussed.

The latest revision of the AECMA 1000D standard is tailored to IETM production, whereas the ATA 2100 standard is not. AECMA 1000D works with data modules as containers for the service information. A data module is identified by a data module code, which specifies meta-data including the type of information and the codification for the part of the equipment being described.

ATA 2100 structures and stores the information in different manuals for service, maintenance, provisioning and other types of information. The aircraft manuals are broken down by chapter, section, subject and to a finer level of granularity, depending on the type of manual. To harmonise the approach for both standards the ATA 2100 data set has also been redefined as documentary units allowing the production of database driven IETM's. Other similarities and differences between the AECMA 1000D and ATA 2100 standards will be highlighted and discussed.

ATA 2100 data

ATA 2100 data sets consist of long linear SGML-instances with a typical chapter, section and subject breakdown. The structure is set up to publish traditional paper documents. However, deeper in the hierarchy richer elements can be discovered. These element names such as <task> or <figure> indicate the functional contents of the element. The function of these elements can be compared to the concept of data modules that AECMA 1000D is using.

The exact structure of the ATA 2100 data differs per document type. At the moment ATA 2100 has more than 10 different document types, such as Aircraft Maintenance Manual, Illustrated Parts Catalog, Electrical Lists Manual, Wiring Diagram Manual and System Description Section.

The ATA elements can be divided into two groups. The first group of elements allows for structuring the information in typical layout oriented blocks such as <para>, <table>, <randlist>, <warning> and so on.

The other group are elements representing or referencing to other entities such as: other data sets, consumables, parts, illustrations, electrical equipment, zones, and so on. Some of these elements are represented as internal (<refint>) or external (<refext>) hyperlink constructions. Others are implicitly defined as elements with normal attributes. Examples of these elements are <con>, <csn>, <zone>, etc.

A special feature in the ATA 2100 data is the use of in-document applicability. On several levels one can define the applicability of the element by specifying an attribute value to be a certain aircraft or equipment number range. This method allows for maintaining the information for different equipment variants in the same source document.

Another special feature is the marking of revisions. ATA 2100 marks revisions with the separate elements: <revst> and <revend>. The DTD does neither restrict the placement nor force the levelling of these elements. The elements are single empty elements that are included at top level and that consequently can be placed at any position in the structure. The author is responsible for levelling the start and end markers.

AECMA 1000D data

AECMA 1000D is based on the concept of organising the data in data modules. Data modules are identified by a data module code .

Note:

^{A data module is uniquely identified by using its logical identifier (data module code) and other information located in the status section (language, issue number, language, ...)}

the model of the equipment the data module concerns,
the relation of the data module to the system breakdown (chapter, section, subject, ...),
the type of information contained in the data module.

AECMA 1000D covers 6 different content models, each with its own structure:

procedural,
descriptive,
illustrated parts data,
fault isolation,
schedule,
air crew information.

Depending on the type of information that is stored in a data module one of the content models is used. The structure for each content model is strictly defined and tailored to the function of the data module.

In the data module references to other data modules, illustrations, parts, tools and materials are made. These references are not actual hyperlinks, but semantically represent a relation to another entity. Examples of these elements are <refdm>, <reftp> and <xref>. To identify the different targets of <xref> an attribute "type" is used with the following allowed values: "TEXT", "FIGURE", "TABLE", "SUPPLY", "SUPEQUIP" and "SPARES"

Main characteristic of data module is that it is self-contained, so it can be stored into databases (Common Source DataBase) and shared between partners. Technically it is also self-contained because each data module is an SGML document.

A publication contains Data Modules extracted from the Common Source DataBase and structural elements (i.e. tables of content, indexes,...). As each data module is an SGML document, a publication is described as a HyTime hyperdocument.

A publication is defined by its content (data module to be included) and its final structure. The structure of the publication is defined in the "publication outline". The data modules are automatically inserted in the publication outline.

AECMA 1000D defines a publication format (a HyTime Document for IETP-DB) enabling the exchange of a publication between partners or between manufacturer and customer. On the other hand, end-user deliverable publications contain presentation (style sheets, interactivity, ...) which is not formally defined, because it is a rapidly evolving area and most of implemented solutions are proprietary.

Both ATA 2100 and AECMA 1000D data contain references to position numbers on an illustration. These references to call-outs are not tagged as SGML elements, but may for instance be recognised by certain semantics such as a number with surrounding brackets.

IETM Requirements

Browser requirements

Dividing the data set into smaller grains of information allows for easy navigation through the information but is not good enough. The users of the IETM expect more. Besides easy navigation and searching the end user and the distributor of the information will expect additional functionality of the application used for the implementation of the IETM.

Additional functionality is for instance:

Ordering of parts:

(On-line) ordering functionality is a logical extension since the parts data is readily available.

Display of warnings, cautions and notes:

Typical for IETM’s is the quick navigation through the information. Important information such as warnings and cautions could thereby easily be missed if no special precautions are taken. The information has to be displayed in such a way that the user cannot overlook it and has to confirm that he read it.

Adding annotations:

In a traditional paper based environment the users often annotate the original information. This functionality is also demanded in an electronic environment. This could be additional information, a remark for his colleague or a notice that the text stated is wrong. These annotations should remain in context when the data is updated.

Applicability of text:

Removing non-applicable parts of the data module during data preparation is an option to simulate applicability handling. However, this requires that for every variant of the equipment a specific data module has to be created. This could result in an explosion of data.

A better solution is to provide applicability mechanisms in the browser. Based on the equipment variant the user has indicated and the applicability coding in the data module, data is displayed or hidden appropriately.

Zooming on illustrations:

Providing the user with both the ability to see the details and to keep an overview of the illustration is a challenge in an IETM. One of the expected functions of an IETM is that the user can zoom and pan on illustrations. Especially vector drawings are very suitable for this purpose.

Besides zooming, the electronic medium also allows cheap distribution of coloured illustrations. Colouring can improve the functionality of the illustrations.

Integration of non-SGML/XML (legacy) data modules:

Although a class IV IETM in theory should based on SGML the need for to integrate data from sub-suppliers or from other systems is high.

Preparing and enriching the data

Requirements to the data preparation

An IETM class IV requires that the information be organised in data modules containing a complete meaningful block of information. For AECMA 1000D data the information has already been divided into data modules because this is part of the standard, however for ATA 2100 data the data must be divided into logical elements such as AMM task elements or IPC figure elements.

Consequence of dividing ATA 2100 source documents into data modules is that originally internal hyperlinks may go out of context and must be converted to external links. This conversion must be handled during the division of source document into data modules.

The user of the IETM will demand more of the data. For example, he wants to know more about:

A consumable: what is it and where can I get it?
A part: is it on stock? Can I repair it? Who made it? Is it an assembly or a subassembly?
A zone: where is this zone?
A tool: how to use this tool?
An IPC item: where is the item number on the illustration?

This requires linking between associated elements and data described in other data modules or even in graphics.

Another requirement of the user is that he wants to see only information applicable for the equipment he is currently working on. This requires that non-applicable data is not shown or is removed during data preparation.

Data preparation tool

Creating an IETM can be a labour-intensive activity depending on the amount of data and the requirements. Especially adding linking information to the required level of functionality, including links to call-outs in illustrations, was expected to be a cumbersome task, if possible at all when done manually. For the first IETM based on ATA 2100 data, the expected amount of hyperlinks was 35.000. Besides this initial effort for the first edition, one would have to redo all the work for the next production when the data set would be updated.

Automation of the production of IETM’s with such a high level of interactivity is the only solution. Therefore it was decided to develop a preparation tool able to generate a complete IETM Class IV set of data modules based on ATA 2100 and/or AECMA 1000D compliant data.

The functions of the preparation tool are:

analysis of the data sets,

division of the data into data modules. In this process existing internal links must be converted to external links if the target anchor is outside the current data module. For these external hyperlinks HyTime is used. In the future, however, the preparation tool could also generate normal HTML links or XLink’s.

creation of additional hyperlinks to enrich the data. The preparation tool can create multiple hyperlink types, for example:

From consumables to the appropriate description in the first chapter of the manual,
From special tools to the appropriate item in the illustrated parts data,
From routine replacement parts to the appropriate item in the illustrated parts data,
From part numbers to their stock information,
From assemblies to their sub assembly and vice versa,
From manufacturer code to the appropriate additional data.
From a wiring diagram to other related wiring diagrams.

In this enrichment process links between hotspots on an illustration and call-outs in the running text are created as well. By automatic analysis of illustrations in text-encoded CGM format or TIFF format, a list of hotspots and their position on the illustration is created. The call-outs in the text are matched with the hotspots in this list and converted to appropriate HyTime elements. This mechanism is also used to create bi-directional hyperlinks between hotspots and items in the illustrated parts data.

The core of the enrichment process is a production database that stores all information about the data modules, its contents and relations to other elements. This database also enables validations of the data. Where the authoring environment typically does not check whether a consumable is encoded correctly, a mistake is now easily detected. Results of these validations are fed back to authors for the next release.

The preparation tools transform the marking of revisions to proper surrounding elements using one element: <rev>, with a start and close tag.

To allow this enrichment of the data, the original set of DTD’s is extended, resulting in a new separate set of publishing DTD’s. In the browser the data modules will be parsed with a publishing DTD. In theory parsing is not required, but as the browser provides this check automatically it is still used.

Last but not least, the preparation tool has to generate one or more structures for the user to access the information. Using electronic media enables multiple access structures on the same data. In our case the tool can generate a hierarchical representation, say tree structure, of all included data modules based on the ATA 2100 breakdown, AECMA 1000D system breakdown and AECMA 1000D information code breakdown.

Results of the project

The project has resulted in three main products:

an off-the-shelf available browser that suits the requirements of class IV IETM’s
a preparation tool to produce IETM’s based on ATA 2100 or AECMA 1000D data and
a hotspot manager to manage hotspot positions on illustrations.

In addition to the products there are experiences to share:

Although SGML steers the structure of the data it does not steer the contents, so authors make typing mistakes. As the enrichment is an automatic process this will result in missing or invalid hyperlinks. To overcome this, the data is analysed before the links are created,

standards do change. The AECMA and ATA 2100 standard are no exception, therefore the preparation process should be flexible and easy to tailor,
it is possible to approach AECMA and ATA 2100 data using the same principles
IETM productions differ per production, therefore the preparation process has to be highly customisable and configurable per production,
data is changed continously, also during production time of an IETM. This demands a preparation process that should be able to be re-run to reflect changes in the source data in the IETM,
browsers and standards will evolve, new standards like XLink and XPointer will be adopted soon. Therefore the output of the preparation tool should be easy to adept to support other formats.

Conclusions and remarks

Producing an IETM is something anybody can do. Producing a good IETM class IV that is accepted by the user however takes more. Next to the first production of IETM, the maintainability of the production is an issue. The approach to automate the production of the IETM from scratch on the basis of an updated set of source data is a feasible and better solution than continuously trying to update the data set of the IETM on the basis of changes in the source information.

The classification by DoD gives an idea about the functionality in a class IV IETM. However, it does not say anything about the interactivity. This could be something for the ATA 2100 and AECMA 1000D committees to define.

Both standards provide a rich encoding standard for their data, allowing the production of advanced IETM’s. As both committees are evaluating or working on the step to XML, some improvements could be made in parallel.

For ATA 2100 this could include changing the revision mechanism which is not very strict and structured. The use of XML will already steer this as the use of inclusions and exclusions is no longer allowed. Both standards should define an interactivity level for IETM’s, for instance by defining which elements can be source anchors for a link to a target anchor.

The main reason to go to XML is of course the fact that free browsers will support the browsing of XML instances. Whether or not this would provide the user with the functionality and interactivity he expects, is an issue that requires evaluation by the producers before they take this step. Users expect more of the data the moment it is electronically available.

One of these aspects is the use of coloured illustrations, which was highly appreciated by the users. The hiding of non-applicable data is another very important one. Not only from a user-friendliness point of view, but also from a liability standpoint of the manufacturer.


XML in Investment Banking		Table of contents		Indexes		A Healthy Revolution