Freescale’s Big IDEA
Freescale began its journey toward XML-based information development, management, and delivery in 2005 when we took an inward look at documentation processes. That look revealed a rather startling metric: technical writers were spending nearly 70 percent of their time doing things other than technical writing! Applying the 5 Whys1 test led to the discovery of two root causes:
- Content was scattered all over the globe in multiple repositories (including local harddrives), making reuse nearly impossible.
- That already scattered content was stored in multiple binary file formats in which presentation was tightly interwoven with the content. This forced technical writers to spend time converting from one file format to another and modifying presentation rather than working on content.
To address these root causes, we created a single, unified repository for all content that would be stored in a presentation-neutral, non-proprietary format. The result is an XML-native repository that “speaks” DITA, SVG, and MathML as its native language and is accessible by any Freescale employee across the globe.
As we dove deeper, we made even more valuable discoveries that reached beyond simply “fixing documentation.” Thus, what started as a solution to a documentation problem quickly grew into a much more pervasive information-management solution that touches processes integral to product development. We call it our
Applying IDEA methods boils down to breaking away from the “throw it over the wall”
approach to documentation and, instead, integrating information development throughout the entire product development life-cycle. A key part of this approach is morphing the 5 Why test into the 5 Where test. When we need to display an information object in a document we ask, “Where do we find that information?” We’ll be pointed to some other information object, and again we’ll ask, “Where do we find that information?” We’ll get another pointer, and we’ll ask the “where” question again…and again…and again. Eventually we arrive at root information. We then work to build a content flow that streams root information directly to consumers, rather than stopping several times along the way. Wherever possible, we also attempt to automate those flows to allow information developers more time to create content rather than moving it around.
Unanticipated value has grown in the ecosystem. Once we’ve created these automated, root-information-based flows for the purposes of delivering documentation, we can leverage that same information in non-documentation flows such as software and hardware development, as well as testing. The entire ecosystem takes on the form of a virtuous circle. From initial market requirements gathering to final manufacturing, we’re striving to make information development the foundation of each step. And we’re seeing remarkable results (see Figure 1).
Figure 1: Information Development Ecosystem
Efficiency and Effectiveness
The old saw says that one must be careful to be effective (doing the right job) while being efficient (doing the job right). Before moving to an XML-based information development ecosystem, improvements in efficiency frequently came with the risk of compromising content accuracy; in other words, ineffective efficiency gains. As part of our IDEA program, we have used the capabilities of XML and centralized content management not only to drive greater efficiency in several key areas but to do so while actually enhancing content accuracy. We have identified and concentrated on several areas of increased efficiency.
Through single-source techniques such as consolidated and parameterized content managed in an XML-optimized CMS rather than in traditional software design repositories, we spend less time having to validate content across multiple copies or versions and across manual data transforms.
The same single-source techniques and a centralized content management system allow us to spend less time locating, tracking, transporting, and aligning multiple copies of content.
Aggressive Reuse Strategies
By analyzing the potential for content reuse across product families and leveraging the automated processing made possible by an XML-optimized CMS, we have greatly reduced production times for family documentation sets. For example, in one particularly complex and extensive product family, we are able to generate more than 30 tailored sets of consistent product documentation, comprising over 60,000 pages of content, using a single-sourced content data set of roughly 3,000 pages. Furthermore, we can generate all of those documentation sets in less than four hours, versus a two to three week process with our legacy system.
Automated Generation of Specialized Views of Content
Specialized data models in a machine-readable format (XML) allow us to develop automated processes for delivering tailored views of content to different internal consumers, increasing content reuse, and extending the value of our information development teams and processes beyond customer document production. Our information development processes are now becoming integral to the design process, and the efficiency gains extend through the entire product development cycle. At the same time, using these processes that operate on a single-source of content simplifies the tasks of ensuring and maintaining content accuracy.
For example, in the industry-specific information domain of register definition2, we are now feeding processed data packages based on the same source used to produce our customer-facing documents to our software enablement teams. These single-sourced data packages allow those teams to deliver product-specific plugins for our software enablement tools two months before customer samples are delivered, instead of the two to four months after customer samples that were required by our legacy process (see Figure 2).
Linkage of External Product Data into Customer Content
Freescale is a major supplier of semiconductors to the automotive industry. To compete in that market, our product development and documentation processes must comply with the emerging ISO 26262 standard “Road vehicles—Functional safety.” Part of complying with this standard is the ability to trace a safety requirement from a requirements database to the location in a customer document that states how the product implementation satisfies that requirement. In our legacy desktop publishing environment, we made several failed attempts to establish this traceability.
With our XML-based system, we have identified and are now implementing requirements traceability that leverages native XML capabilities.
Keeping the Right Focus
For decades Freescale (as Motorola’s semiconductor group prior to 2004) has employed a multi-tiered customer relationship strategy that includes:
- Direct account teams for our largest customers
- Certified third-party distribution teams for our 2nd tier customers
- A “low-touch” self-service web site for 3rd tier customers
- A focused university relationship program for the engineers of tomorrow
The extremely broad customer base that we support worldwide has always put pressure on information development teams to meet a wide variety of documentation needs throughout a product life cycle. Product proposals grow into product specifications, which grow into reference manuals. Enablement teams within Freescale require early access to content and work in close partnership with customers through the entire content lifecycle to ensure that we are delivering the information they need to succeed.
IDEA methods bring a much more customer-driven focus on how we present and deliver information. As Automotive Applications
and Information Development Manager Mark Robson notes, “Information development is now being recognized for its business impact. As products increase in complexity, it is imperative that creativity and initiative are applied to the documentation process in order to enable the organization to deliver a combination of a compelling product with the ability for the customer to get the most out of it.” A groundbreaking piece of silicon does nothing for the bottom line until the customer knows how to use it!
By introducing efficiency into the way information developers work, the ecosystem enables the creation of more user content, which fills in the gaps between what a product is and how a customer can use it. A recent survey of Freescale employees and customers revealed that this is one of the top requests for documentation. It also put a spotlight on the preferred deliverable formats. While online PDF files are still dominant, customers are beginning to show a strong interest in interactive HTML and online databases. Within the ecosystem—particularly with the launch of a product family infocenter next year—these customer requests are being realized as innovative deliverables.
The ecosystem also gives the information developers more time to collaborate directly with customers. Some examples of such collaborations are
- Customer reviews of early documents (through non-disclosure agreements) directly drive content improvement.
- Direct customer meetings that focus on information deliverables tell us where to focus our efforts.
- Customer satisfaction surveys clearly identify areas where we’re doing well and where we need to improve.
- Co-development of information development methods in support of standards such as ISO 26262.
Beyond applying the IDEA to Freescale’s existing means of enablement, we are also seeking to create a new way of interacting with the customer. The Reality Check program extends the reach of the ecosystem so that we can collaborate with our internal and external customers to create a better user experience. The key focus of the program is a web portal that provides our customers with a feedback-rich environment to explore and review new features regarding content creation. We envision the Reality Check program as a means to
- Gather customer requirements
- Solicit and elicit customer feedback
- Listen continuously via service centers, surveys, emails, blogs, and forums to gain unsolicited attitudes and perceptions
- Exploit Web 2.0 applications that encourage information sharing to gain further insights into customer needs and opinions
- Inject ourselves into opportunities for direct contact such as trade shows, customer visits, events, and so on to gain a true understanding of the customer’s side of the equation
- Track relationships to evolve detractors and passives into promoters in Net Promoter Score surveys
- Monitor interactions and discover best practices for enhancing documentation
- Disseminate findings and infuse each information development project with viable insights that are actionable
This synergistic approach sharpens our customer focus, engages the customer in future information design decisions, and secures customer buy-in before the implementation of any changes.
Sharing the Wealth
One of the most encouraging discoveries along this journey has been that our IDEA is highly transferable. Where we traditionally had functional “silos” that didn’t share methods and core technologies, we now find that the information development community is bringing its expertise to product development teams, and in turn they’re learning a great deal from those other teams. It’s happening throughout the company as new content management and eMedia technologies broaden the ecosystem advantage.
Exposing the Core Value of an Information Development Team
The Semiconductor Information Design Subcommittee3 (SIDSC) was created to strengthen the ecosystem because other industry standards, designed to enable circuit design automation, had not provided better documentation support over the years. The creation of SIDSC was led by representatives from global information development teams for use in documentation. After some time, it became clear that adding a few more semantics could make it the most robust, publicly known register data model in the industry. So, those semantics were added and the standard is quickly gaining momentum.
The moral of the story is that information developers do not have to wait for industry standards to emerge from the design community to solve their problems. They have the power to build what they need. And in some cases, what they build will become the industry standard.
Information developers are not just experts in the content that appears in documentation. They are, as their title suggests, experts in information. The new role of the team is to leverage that expertise in a way that brings exponential value to the business as a whole.
Transferable Data Models
The leveraging of expertise is clearly transferable across organizations. More practically, we’re finding that all of our tools and technologies are transferable. The ecosystem uses standard DITA types for most content:
- OASIS DITA SIDSC for register data
- Other specialized DITA types for pin information (which generates both a table view and a graphical “pin-out” diagram in SVG)
With the right analysis, anybody can trace information back to a root source. If that root information is highly structured, anybody with DITA specialization expertise can devise a data model that makes the content more relevant to more consumers and enables intelligent business-to-business porting to different formats as well as better metadata-based content management.
In addition to transferring knowledge between information and product development teams, we’re finding that IDEA methodology accelerates innovation. This happens as cross-functional teams within the company engage in problem solving while encouraging them to subscribe to an integrated information flow rather than serving as mere providers or consumers of content. In addition, by proving the business value of a DITA specialization, Freescale leverages the “wisdom of the crowd” embodied in the DITA TC and the SIDSC to lead the way. This has brought together a group of competing companies and organizations to share their common pain points and collaborate on a common solution that helps all parties. Innovation shines in several areas.
Extending the Value of Information Development Beyond Documentation
We have extended the value and scope offered by traditional information development by playing a leading role in the definition and eventual adoption by OASIS of the SIDSC specification. Through the implementation of this specialization, we have enabled a machine-processable transfer of information between functional groups involved in semiconductor product development. The innovation here is that many of these processes were previously performed by hand, such as verification and documentation. This new way of transferring information has
- Improved time to market while at the same time improving quality
- Decreased the opportunities for errors to be introduced into content and into the design phase of product development
- Provided a normalized definition of our intellectual property
- Provided a platform for a community-prescribed, open-adoption pattern to information management
Improving Time to Market While Improving Quality
Fostering the IDEA has leveraged reuse and re-purposing of content to a level not previously achievable with legacy processes. Eliminating redundancy of design information has yielded shorter cycle times for review and correction. Shorter cycle times in semiconductor product development translates directly to faster delivery of Freescale products to market. In addition, the shorter cycle has not come at the expense of quality. On the contrary, it comes with an increase in quality, both in terms of reduced errors (accuracy of documentation) and greater customer confidence in the documentation.
A serendipitous result of the IDEA is that downstream consumers who previously relied on human-generated design information (in the form of PDF files or other traditional documentation formats) are now able to consume machine-processable design information.
This means that automation can be applied in a way that was not possible previously. Processes that used to take weeks now take hours. Freescale’s implementation of the IDEA has allowed our information development teams to drive innovation in these areas:
- Software tools used by our customers for embedded system application development
- Internal verification tools used to verify the integrity of product design
- Electronic Design Automation (EDA) tools used by designers to design and analyze semiconductor product design
- Application notes and errata reports used by customers and field engineers to understand device-specific implementation
Improving the cycle time of each phase of product development has resulted in the delivery of product to market never before seen in the semiconductor industry.
Increasing Technical Accuracy of Documentation and Design Specifications
The IDEA improves technical accuracy in both technical documentation and in design specifications. Freescale’s implementation of the SIDSC register data DITA specialization has eliminated opportunities to introduce errors and enabled mechanisms that facilitate verification of product design information.
In addition, because of the integration of the information flow across functional groups, requirements tracking and traceability information is now available to both upstream and downstream participants in the information flow. For example, a customer requirement that is introduced at the time of design can be associated with metadata that follows not only the intellectual property in the product design that implements the customer requirement, but also all other phases of the product definition such as verification and documentation.
Standardizing and Normalizing Intellectual Property
There is a competitive advantage to focusing on the ecosystem of information created by providers of content and the consumers of content. The ecosystem includes different functional groups within the company as well as joint-development partners and customers. The key to the competitive advantage is in enabling different participants in the ecosystem to share design assets in a standard format.
Providing an Open Platform: Community owned and community managed
The platform that we have implemented uses an approach to information management that is community-prescribed and based on an open adoption pattern. Community-prescribed
means that participants in the community (both providers and consumers of information) have a voice in the creation and extension of the platform. An open adoption pattern means that the platform provides mechanisms for adopting the solutions that work, regardless of where the solution originates. Furthermore, solutions that work for one member of the community can immediately be distributed to other members of the community. Weekly ‘Ask the Experts’ sessions ensure the ecosystem’s free flow of best practices and provide everyone from the first-time user to the core team member the opportunity to learn something new.
Perhaps most innovative of all is that anyone in the community can contribute a solution to the platform, regardless of title, position, or organizational structure. The result is that Freescale is now tapping the unique talents of members of the community whose contributions would otherwise have been overlooked. Fostering the IDEA is by definition fostering a spirit of innovation.
Transforming the Culture
The IDEA has grown from a single summer internship pilot project to encompass multiple business groups with stakeholders reaching the top of the corporate org chart. Freescale’s transition from focusing on hardware to driving solutions has spurred a dramatic shift in the way that documentation is viewed across the company. Now recognized as vital to enabling the creation of solutions as well as customers’ application of them, we believe that IDEA contains the best practices for documentation within Freescale and beyond.
In an April memo distributed throughout the company, two VPs and a Senior Member of
Technical Staff pointed to the improvement of documentation through the ecosystem as a key to success. “Everyone has a role in moving Freescale toward flawless documentation…