When experts or professionals deal with problems, they tend to approach the problem based on historical experience, current observations and future trends. They usually recall similar experiences and draw on best practices and lessons learned. For most of us in information technology, the concepts of reuse are not new, but few of us pay homage to Christopher Alexander and his original work on utilizing patterns in architecture. The object oriented crowd picked up on patterns and changed their industry. A recent book, The Design of Sites (Van Duyne, Landay, Hong, 2002), describes a collection of patterns for building Web sites. The book begins with a short chapter on the foundations of good design, which provides principles and standards and focuses on customer-centric goals. The core of the book is the 12 design categories based on real examples, which are described in detail. Is it possible to create patterns for enterprise metadata? The first step is to establish a collection of categories where we can classify the patterns of the organization. One such collection might include structure patterns, design patterns, process patterns, content patterns, service patterns and architecture patterns.
Structure patterns focus on the core data structures and standards defined by the data architecture organization. As the architects of metadata, we are responsible for defining and implementing metamodel standards such as Reusable Asset Specification (RAS), Object Management Group (OMG) or Universal Description, Discovery and Integration (UDDI). These standards should be cataloged and made readily available to the technology community. While core asset metamodels are critical, integration patterns such as the Dublin Core are also important. Many data architects focus on building domains for specific fields such as race, gender, months, days of the week, etc. These standards, hence patterns, are critical to integrating data across systems and should be captured by a central repository. Other structures patterns such as Records Information Management (RIM) Sarbanes-Oxley (SOX) and IT Infrastructure Library (ITIL) are emerging and adding to the portfolio.
Design patterns focus on the human computer interaction of the repository or registry. Clearly, our industry has a way to go to implement design patterns which are critical to the success of any project. We learned a long time ago, that making life easy is more important to the vast majority of users than adding functionality. There are various schools of thought on which design elements make a successful Web site. Scanlon, Schroeder, Snyder and Spool (1998) collected qualitative and quantitative data on key design factors, which included: searching, content, text links, images, link navigation, page layout, readability, graphics and user's knowledge. Each of these design elements makes an important contribution to a successful repository.
Repositories are built to provide metadata information and provide services to both the producer and consumer of the information. Experts indicate that usability is about making sure that the average person can use the site as intended. Well-chosen names, layout of the page, text, graphics and navigation structure should all come together to create instantaneous recognition (Krug, 2000). There are a variety of Web design elements that can have a positive impact on a Web site's image, effectiveness and trustworthiness. Design elements such as carefully selected images, clean and clear layout, careful typography and a solid use of color can create an effective site (repository). In addition, a solid navigation structure and continuity in design can provide the user with the control and access required within an e-commerce interface (Andres, 1999). Although, design elements may take on the form of a visual cue, the true value comes from a combination presentation, structure and interactivity. A solid repository is a collaboration of design components, content (metadata), usability patterns and back-end systems that integrate the core business processes. In our library of patterns, we focus on navigation, page layout, images, styles, search components, writing techniques and trust/relationship patterns. This consistency of design enables us to integrate the disparate repositories into a single metadata portal for the entire technology community.
Process patterns focus on specific business processes within the organization. Basically, we view metadata as a business and our processes are our lifeline of existence. Business processes sound more like jargon than something that is critical to the success of metadata. However, the reality is that processes allow us to serve more customers with fewer resources. Processes allow us to move toward a self-service model; simplification is the key. Processes that are clearly defined, easy to grasp and automated benefit everyone up and down the value chain. Processes can be classified by the following: workflow, subscriptions, consumption services, submission services, procurement, impact analysis, reuse, metrics and data quality. Think McDonalds: It doesn't matter where you go, the process for cooking fries is the same.
Content patterns focus on the different types of content accessible to the end user. For example, a repository site may have additional content elements such as who to call for training or frequently asked questions (FAQ). These components represent subject-type content categories. In the metadata space, I can imagine a pattern for the different documents that support the data structures or logical model patterns. Any pattern that focuses on the type or context of the content can be considered a content pattern. One of the best methods of consolidating or publishing the content patterns is the sitemap. The sitemap provides a high-level view of the content, workflow and the overall customer experience for the repository.
While content patterns focused on the self-service information, service patterns focus on the specific business functions within the metadata space. The most obvious patterns are workflow and patterns that correspond to the implementation of business processes. More specifically, we can create patterns for alerts and subscription services for asset notifications. We can create patterns for adding assets to the repository as in a new logical model or data set. Patterns can be built for data extraction, data quality and procurement of services. In the basic definition of metadata management, we must be able to clearly define our product, services and solutions in a way that is easily understood by the average customer. Are service patterns the same as process patterns? Technically they are; however, because services are so important to our business model, they are broken out into a second group to emphasize the importance of the service value delivered.
The final collection focuses on the patterns within the data, application, technology and functional architectures. While there may be some overlap with the other classifications, the architecture patterns tend to focus on the traditional delivery of the information technology organization. In most cases, architecture patterns are more internally designed versus the external facing ones previously described. Enterprise architecture planning is the process of defining architectures (data, application, technology and functional) for the use of information in support of the business and the plan for implementing those standards (Spewak, 1993). Data architecture focuses on the data quality, data management, data content, data usage, modeling, storage, and traditional metadata management. Technical architectures review the hardware, software and vendor relationships while the functional architecture documents the business processes. The final architecture is the application architecture, which works as a conduit between the functional and technical specifications. Architectures define the rules of the game within the corporate environment, and these rules can make or break the Web application implementation.
Clearly, I'm evoking the element of reuse in discussing patterns. We want to deliver products and services in a consistent, efficient and timely manner. Patterns are not new, and I'll bet everyone has their own method or process storing, archiving, locating and using patterns. The first step in creating the documentation of a pattern is to describe the pattern on a single page - no more and no less. The key is to ensure that you can define and explain the pattern as concisely as possible. The one-page requirement ensures that patterns are as simple as possible, but it has linkages to more specific information as needed. The page will contain a title as well as a classification scheme that corresponds to the needs of the business. In addition, each pattern will present the problem the pattern is addressing, the summary of a solution, an example of application, and benefits and consequences of application. Additional information can be linked into the document for step-by-step instructions, visual examples or detailed documentation. We shouldn't lose site of the goal: reuse. If we design patterns that are complex and difficult to understand, people won't reuse them which will destroy the business incentives.
Patterns are not just for programmers and architects - each and every member of the data resource management organization can benefit from an inventory of solutions. One of the misnomers is that a pattern should be groundbreaking, patentable or something that no one else has ever thought about. That type of thinking will force you to miss a great majority of the value that patterns can bring to the table. Simple patterns can be very effective; I would rather have 100 simple patterns than five complex ones. Don't get caught into this type of thinking; value and reuse have no boundaries.
Register or login for access to this item and much more
All Information Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access