November 1999
Web DNA: XML Builds Powerful Publications
By Liz Levy
They’re two very different companies with two different challenges, but they’ve
both found an answer in XML (eXtensible Markup Language). What they have in common is a need to
organize and add structure to their documents with an eye toward delivering the content across
multiple media including the Web.
Read on to find out how Practitioners Publishing company plans to customize and reuse its financial
reference material for delivery via print, the Web and CD-ROM. And there’s Guidant, a leading
maker of life-saving pacemakers and defibrillators. Using XML, this manufacturer is creating
structured documents that can be dynamically composed for each product they manufacture.
Embracing Web Delivery
Practitioners Publishing Company (PPC) offers a variety of reference materials used by CPAs and
financial planners. There are 300 employees in PPC’s office in Fort Worth, TX, where they
publish more than 125 titles a year averaging more than 1,000 pages in length.
Embracing advanced multimedia opportunities, PPC’s goal is to offer customized publications
to its customers over the Web. By accessing content online, customers will be able to select the
portions of PPC’s products that pertain to their specific needs. PPC also wants to facilitate
the reuse of content in multiple publications delivered through print, the Web and CD-ROM.
“We eventually want 50% of our revenue to come from Web-based products,” says Garry
Gellett, program manager at PPC. “A lot of our material changes constantly due to new
legislative requirements. The Web offers more timely delivery so our customers can get the most
up to date information and don’t have to wait for the CD or print versions.”
Most of PPC’s documents were published in print first and then on CD-ROM. Until recently,
only a few of their products were made available online. Online delivery was accomplished with a
proprietary image viewer. Gellett says PPC had considered using a standardized markup language
— namely SGML — for a number of years, but they found it too rigid for their needs.
“There was nothing that suited our requirements for reuse and timely delivery,” Gellett
explains. “We watched XML come on the scene and saw it gain support from companies like
Microsoft and Netscape. It had the extensibility and flexibility we were looking for.”
XML facilitates the reuse of content by storing information at an element level rather than a
document level. This fit PPC because some of their products have the same sections and
subsections but they don’t have the same chapters.
“With XML we can author the information once and produce it many times without additional
work,” says Gellett. “There is no time wasted looking for the right pieces of
information and cutting and pasting.”
PPC often licenses their content to other publishers. These clients, such as Lexis/Nexis, are
information aggregators that act as a one-stop shop for customers to purchase financial and
accounting information.
“We were losing licensing opportunities because many of these companies required SGML to
exchange content,” says Gellett. “We also have sister companies that were unable to
share content with us.”
XML, which is compatible with SGML, held forth an answer to PPC’s problem. Late last year,
the company began implementing BladeRunner, an XML-based content management solution from
Interleaf (www. xmlecontent.com). The system includes tools for content creation, management
and publishing.
PPC took part in an early adopter program of the then just-released BladeRunner system, working
closely with Interleaf’s software engineers on the application development in two phases.
The first phase began last July, and it entailed taking one of PPC’s published products,
creating a DTD for that document and importing it into an XML repository.
BladeRunner supports the use of Microsoft Word and other XML authoring tools as an interface to
create structured documents. It validates content against a DTD, which is created graphically using
BladeRunner, and parses documents into reusable elements. BladeRunner also supports XSL
(eXtensible Style Language) style sheets, which render documents for different publishing mediums.
PPC had been using Interleaf’s document editing software (now Interleaf 7) for more than five
years, so they were accustomed to authoring structured content. They continue to use this as their
authoring tool along with Microsoft Word.
“We sat down with a very simple document and looked at where we wanted rigid authoring rules
and where we didn’t care as much,” explains Gellett. “For example, in the DTD,
wherever you start a new chapter, the author has to write a chapter title — otherwise it
would not be valid XML. Another example is the sections of the documents that have headings and
subheading within them. In our DTD, we created a rule that you couldn’t have a subhead without
the section head above it.”
When using XML, you name the elements intelligently by what they are; the style is applied later.
This makes it easier to apply different styles to the same element.
“We had to get away from naming our elements based on what style they had,” says Gellett,
explaining that tags such as “double indent 15 point” had to be replaced with
“subhead.” “Even in our repository, we had documents named for the style they used
when printed. The problem was, different styles were applied for Web and CD-ROM output.”
With XML, XSL style sheets are used to render versions for print, CD-ROM and the Web. PPC also uses
XML for product catalogues, which can also be easily updated and recast for different media.
In the second phase of the project, PPC began rolling out the system to other products, bringing the c
ompany’s most popular material to the Web first. At press time, PPC was still in its early
stages of development with the BladeRunner system. They had 12 products available on the Web and
were working on publishing a total of 35 products online by the end of September.
“BladeRunner will allow us to publish documents online immediately,” says Gellett.
“Once the authors check documents into the XML repository, we can set up everything in the
production process to happen automatically.”
PPC expects to reach this advanced stage within two to three years. By doing more business over the
Web, the company will save overhead costs on things like shipping. More importantly, the company will
be able to provide better service and a better product, says Gellett.
“We are early in the process of developing an application that is based on relatively new
technology,” he says. “It’s not a small effort, but we feel it is where we need to
be as we see more demand for products on the Web. As customers become more demanding about timeliness
and customization, we are putting a system in place that will allow us to serve things quickly
and in different ways.”
Putting Flexibility On the Line
Guidant (Indianapolis, IN) is a $1.9 billion, 6,000-employee manufacturer of cardiovascular products.
The company’s Cardiac Rhythm Management Group is a market leader in defibrillator systems,
which treat life-threatening rapid heart rhythms. Based in St. Paul, MN, the group also produces
pacemakers and other products, all of which require thousands of pages of documentation to support
manufacturing processes.
Until a year ago, the Cardiac Rhythm group was completely reliant on paper for distributing
manufacturing instructions to assembly line workers. Because of the sterile conditions required
on the line, the documents were kept in laminated folders and were updated manually. The group
amassed a total of 1,300 to 1,400 paper documents containing about 15,000 pages each.
In 1998 the group began investigating XML as a solution, and after a year of research and system
assessment they began implementing their current system.
“Even now, only a couple dozen
documents are online,” says Craig Benson, project coordinator for Guidant. “We will
slowly add to this number over the next few months and plan to convert all of the documents in
2000.”
“We create the manufacturing instructions for our products in two different ways,”
explains Benson. “Most of the documents are process specific. This requires using large
tables in the document to account for the numerous products that can be built by a single process.
“There are also product-specific documents, which are hard to manage and update, and can lead to
redundancy,” Benson explains.
The Guidant group decided to rewrite all of their manuals in XML (eXtensible Markup Language) to
improve document structure, organization and creation. “It is a one-time effort that will
improve our system overall,” says Benson. “Putting XML around poorly structured documents
isn’t enough. Garbage in is garbage out no matter if it’s in XML or not.”
A year ago, the group chose Epic, an XML-based document authoring and publishing solution from
Arbortext (www. arbortext. com). Epic allows them to create structured documents that can be
dynamically composed for each product the company manufactures.
Guidant’s document processes now start with the manufacturing engineers, who author documents
in Epic, which provides a WYSIWYG authoring environment. Epic presents customized templates for the
documents based on DTDs (Document Type Definitions) that are pre-defined by Guidant.
“This gives [engineers] a solid framework to begin a document and takes out a lot of the
guesswork,” says Benson. “These are complex documents for complex products. Rewriting
the documents is forcing us to find better, more efficient ways to convey the information.”
Guidant’s complex documents need to be distributed to assembly line workers in two overseas
plants, in Ireland and Puerto Rico, as well as in St. Paul. These workers are not engineers, so
the documents need to be as simple as possible for them to understand and avoid errors. Updates
to documents are made immediately, so that each plant has the most current information.
“There is a potential for errors that we are working to eliminate,” says Benson.
“We had checks built into the process to find errors in the [old] system, but now we can
eliminate most of them before they happen.”
The Cardiac Rhythm group plans to put all of its manufacturing documents online. Users will be
able to access the documents via a custom Web site on Guidant’s corporate intranet, thereby
making the documents available to everyone.
The group plans to convert all of documents at a constant rate throughout 2000. This means
putting about 100 documents online per month, as well as training about 100 users per month.
They also have to set up the PCs on the manufacturing floor for assembly line users to access
documents.
Thirty-seven documents are already online, and a total of 50 are expected to be
posted by the end of the year. There are about 1,000 line operators in the group, and about 4%
of them will be on the system by the end of the year.
“There is definitely something to gain by making more information more accessible to
everyone,” says Benson. “Users will not need to leave their workstations to find a
needed document. By eliminating paper we are also minimizing the number of copies of each
document and improving maintenance.”
The biggest technical challenge, he says, is writing
the documents in XML and then making them available online. Rather than just publishing the
XML data as HTML, Guidant is using Epic to generate ASP (Active Server Pages) pages that can
incorporate all of the queries for the process-specific documents. The documents are presented to
users in a standard Web browser as HTML pages, yet users aren’t aware of all the processes
that are taking place on the backend that keep the data dynamic.
“Both our engineers and technical writers are being trained to understand the functionality
of XML,” says Benson. “Our technical writers are being trained further on how to tag
the documents. With six hours of training, the writers can do basic XML tagging and create ASP
pages. Within 10-12 hours, they are confident enough to do more advanced work on their own.”
As sophisticated and intimidating as XML sometimes sounds, Benson says his technical writers
were able to grasp the technology very quickly.
“Information tagged in XML is more dynamic and powerful,” he explains. “We can
trigger actions such as inserting links within documents to activate other software programs.
Our workers need to know what documents they need for a specific product. We plan to assign a
numeric identifier for each product so we can automatically assemble all of the right document
components.”
The system costs for the Guidant group are about $500,000, which includes all of the customization,
consulting, hardware and software. It does not include rollout costs for the company in 2000,
nor the early research they did in 1998. It is a big investment that the company doesn’t
expect to pay off right away.
“We have to go overboard because of the nature of the products and our need to comply with
agencies like the FDA,” explains Benson. “We estimate that if we get just a 2-3%
production gain, the system will pay for itself in about three years. The feedback from the
[assembly line] operators is extremely positive. They are surprisingly eager to use PCs over
paper, and they are behind the project.”
