Delivering the Goods: XML Used at Freightliner

1. Introduction

A manufacturer of large, heavy duty trucks faces a number of business challenges when it comes to information. First, there are regulations. Trucks that weigh over 33,000 pounds must meet weight and size regulations for each of the jurisdictions the truck will be used in. Each of the United States has its own regulations. In addition, each country has regulations. Where a truck will be used determines the regulations that will apply.

In addition, the type of operation the truck will be used for determines the regulations that apply.

At the time the truck is ordered, information on component weights and prices, cargo weights, and the implications of these for the region the truck will be operated in must be known and factored in to the truck configuration. As desirable as a particular option may be, if it causes the truck to violate a weight limitation, the option must be foregone, or something else given up. Having complete and accurate information available when the initial order and configuration are made is essential to selling these large rigs

Similarly, when one of these large vehicles is delivered, a host of information accompanies it. Owner's information is an essential part of the delivery. As the truck is used on the road, repair and service information is needed by those who work on it.

The complexity and volume of the information alone create challenges, but in addition there is the issue of dealing with increasing amounts of information as new businesses are acquired or new product lines are introduced. While the product line continues to be differentiated and new options are added, staffing levels are held constant and the expectation is that existing staff and current process will handle the growing number of documents.

Further challenge lies in the new markets sought both to the north in Canada and to the south in Mexico and Latin America. Documents must be translated into Spanish and French to be usable in these markets and these translated versions must be as complete and accurate and timely as the English versions.

For information professionals faced with these business pressures, ways in which efficiencies can be gained, costs saved, and previously manual tasks automated become essential to success both at the company level and within the departments that create, manage, and deliver the information on which the company relies.

This case study focuses on two departments within a large truck manufacturer, each of which is engaged in projects that introduce structured mark-up using XML and SGML, content and workflow management, and automation of previously manual tasks to achieve immediate payback in cost savings and efficiencies. In addition, these two projects, today separated and focused on immediate gains, lay the foundation for more efficient and automated information flow within the whole company and beyond - to the suppliers who provide the truck components, to other manufacturers who may need to share information, to the dealers who sell the trucks, and to the customers who buy them.

2. So You Want to Buy a Truck. . .

Not everyone has the need or the opportunity to buy a Class 8 vehicle like a heavy-duty trailer, a bus, or a fire engine. If you did you would be concerned about its weight bearing capacity, how much it weighed in and of itself, and what kind of options you really needed to engage in the kind of work you intended to put it to. That information is in the hands of the sales personnel in the form of large binders called sales handbooks, configuration guides, or simply “data books”. These documents come to them from the vehicle manufacturer, and the responsible group within the vehicle manufacturer is a marketing department, whose job it is to ensure the databooks are complete, accurate, and usable.

The data books contain the specifications for, and the price and weight of, every conceivable part of the truck: engine, engine equipment, transmission, axles, suspension systems, brake system, trailer connections, cab interiors and exteriors, tires, chassis, wheel bases and frames, instruments, etc. You can create the vehicle of your dreams, subject to your budget constraints and the regulatory weight regulations for the region in which you will run it.

The information on prices and weights comes from a database maintained and updated by another department: the pricing group. Each item in the databooks ultimately resolves to a datacode, an information element that contains the core description. Within each datacode there can be options that a buyer may select or not.

Datacodes appear in more than one model line, and therefore there is a tremendous efficiency in writing each datacode once and using it across the various model lines. But there is a problem: not all options apply across all model lines. Which options are relevant depends on which model line the datacode appears in. There needs to be a way to write the core information once, store and reuse it as needed, but at the same time have a way to automatically “contextualize” the datacode - that is, to have only the options appear that are relevant to the model being described.

The method by which the single-source datacode will be achieved is through the use of XML and its ability to tag data and the attributes of data. Specialists within the marketing department will use the XMetaL editor from SoftQuad to mark up the datacodes and in the mark-up express the conditions that determine when options apply.

The marketing specialist will store the marked-up datacodes in a content repository, Content@ from XyEnterprise. The Content@ system will initiate a new datacode when the marketing group determines that a new condition must be documented.

In addition to conditionalizing each datacode, the marketing group must ensure that the pricing and weight information for each component is current at the time of publication. The manufacturers and suppliers of these components do change the pricing and weights on a regular basis. Once a generically marked up and conditionalized datacode has been created and stored in the repository, the value of using it is realized in using it over and over again in the datasheets for different models.

In order to ensure that the price and weight information in the datacode is current, each time the datacode is used in a new publication, the Content@ repository will communicate with the pricing and weights repository and the current values for prices and weights for that datacode will be used in the printed version automatically.

Whenever a new book is published from Content@, marketing supplies the criteria for the book, the selection criteria are applied, the datacodes are sequenced, and the dynamic data are resolved to ensure the latest, most up-to-date information is used. The final XML source is passed to the XML Professional Publisher (XPP) from XyEnterprise. XPP creates the formatted pages and produces a PDF file, which goes out on a CD or is printed to create the pages for the binders.

This project is currently underway, so it is too early to fully survey the end results of the effort. It is also a first step in a larger process. Having created a managed repository of XML components, the marketing department is looking forward to other uses of the information. A Web application could be built using this repository to deliver the static information as it currently exists. Or, a dynamic configuration tool could use this information to assist in the decision-making process. Given the accessibility of the information in an automated and controlled XML form, delivery of customized information, associated with each particular order will be possible. And, the ability to manage the translation of the datacodes into French and Spanish will bring a real payback.

3. Who Do You Call When the Truck Breaks Down?

The owners of large vehicles depend upon service providers to know how to maintain and repair their fleets. Service providers rely upon the manufacturers to supply them the information they need to do their jobs.

Within the manufuacturer's organization there is a department responsible for technical service information whose job it is to author, update, and maintain the flow of service-related documents that keeps the large vehicles on the road. This group is responsible for producing service bulletins, service manuals, maintenance manuals, warranty documents, driver's manuals, dealer and direct ship manuals, field service documents, recall bulletins, and component parts catalogs.

The great variety of the document types alone suggests the labor needed to keep the content current, accurate, and complete. Writers and editors must be free to think about the content, not about the form and presentation of the information.

In order to keep up with the volume of this service information, any time a piece of information can be reused (a procedure in a service bulletin may also need to go in a service manual, for instance), ways to do so must be found.

In order to meet company time to market goals, the information must be ready in French and Spanish as close to the availability date in English as can be made to happen.

“Lights out publishing” was the watchword when this department undertook to build on their experience with structured mark-up using SGML and introduce a content management and publishing system that brought the ability to reuse components, assemble documents from a repository of pieces, and publish them (to paper or PDF) in a batch manner.

With a content management system that stores each component with a wealth of metadata, and the ability to search for particular words in the content of the data, an area expert can begin the documentation of a new subsystem by finding the relevant components that described the procedures for the last model. Many of these components need not change, since variations from one model to the next can be relatively small. It is straight forward to use them as the basis for the new system and evolve them as necessary.

Managing the workflow was a key business driver in adopting a content management system. In the past, while the group authored in SGML, it was possible for two people to edit the same component simultaneously. Resolving the differences and determining what the “real” document would say was a manual process. With a content management system, a lock is placed each time a component is checked out, preventing simultaneous edits.

Ensuring that the information get adequate review and formal approval by departmental representatives and subject matter experts was a key concern. With automated workflow today, documents can be routed from desktop to desktop within or across departments, creating a formal trail. In the future, interest is strong in extending the review process outside the enterprise and including an electronic signature for sign-off.

To avoid manual reworking of data at the publishing phase, the content management system automatically checks that each document is valid and parsable each time it is posted back to the repository. Such a step ensures that at publish time there are no unpleasant surprises.

Another automated aspect of the controlled workflow is the ability to manage the translation process. Whenever a new project starts, a synchronization check occurs to see if French and/or Spanish versions exist and the related French and Spanish components are checked to be sure they are at the same revision level as the English. A translation project is started automatically, if needed, and the translation process goes on during the information revision process. When the revised document is checked in, only the changed modules are sent on to the translation group, saving both time and effort.

Translation is done partly by in-house translators using Trados and partly done using outside vendors. The workflow in the content management system can support either model.

The content management system also manages the links within documents for tables, figures, footnotes, and cross-references.

Publishing options from the content management system include integration with Enigma's DynaText for CDROM and some Web delivery, producing an XML source from the SGML source for Web delivery, and producing printable versions through XyEnterprise's XML Professional Publisher.

Future plans focus on further integration with other systems within the enterprise and, over the long term, ability to exchange information with other manufacturers.

4. Conclusion

These two departments within a large enterprise are individually investing in and laying the foundation for greater integration of the flow of information throughout the enterprise and, eventually, beyond company boundaries to include suppliers, customers, and even, perhaps, competitors. By choosing XML and following best practices related to standards and tool selection, they have positioned themselves and their company for greater efficiency, increased flexibility, new opportunities to creatively capitalize on information assets, and, in the end to contribute to the bottom line.

Design and Development by deepX Ltd.