Sequent Computer Systems is the leader in Intel-based data center solutions. We provide our customers with complete data center solutions, integrating hardware and software products with consulting services and third-party partner products. It's critical, then, that our own data center operations model everything we try to achieve for our customers.
In our internal development efforts, the Worldwide Data Management department at Sequent uses a CASE methodology and the Oracle Repository to speed overall development efforts. We use Publisher/2000 from Solutron to leverage the information in the repository and create comprehensive, automated system documentation.
Sequent Computer Systems is a leading provider of Intel-based UNIX and NT solutions for the data center and a pioneer in SMP and NUMA processing. In 1997, Sequent was the world's fastest-growing vendor of servers priced between $100K and $1 million.
The Worldwide Data Management department within Sequent coordinates the development initiatives that run the business. For needs that can be met with off-the-shelf products, we use Oracle applications wherever possible. We're one of Oracle's longest-running customers for those products. But for those applications we build in-house, we use Oracle Developer to implement CASE (Computer Aided Software Engineering) methodologies and the Oracle Repository to store the reusable objects, data and process definitions created from our development efforts.
Not surprisingly, we use Sequent hardware for our internal production and development systems. The repository runs on a Sequent SMP system running DYNIX/ptx (the Sequent UNIX implementation). Developers work on Windows-based Sequent systems.
Oracle Designer is a business and application modeling tool that generates complete applications from visual models. An important part of the package is the shared repository that it creates of reusable objects, business rules, data definitions, etc. The repository becomes a rich meta data store for the applications in development. By using the repository to store all new development models, we achieve the benefits of having a central store of meta data for our applications.
The challenge is accessing and using that meta data for systems documentation and information sharing. Sequent uses Publisher/2000 from Solutron, Inc., to extract information from the repository in a variety of different document formats. Publisher/2000 provides automatic, dynamic document generation from the repository. It gives users and developers alike the information they need about systems, while shielding the complexity of the repository itself.
For example, the repository holds all definitions of terms and processes in the systems in development (as well as those we have reverse-engineered into the repository). We use Publisher/2000 to generate process maps, which describe all of the meta data involved with a specific process.
Sequent is in the process of developing several new transaction processing systems using the repository. At the same time, the data warehouse group uses Oracle Designer to model star schema designs and uses the information in the repository to aid in system and metrics design. The centralized repository makes the process and data information available across the enterprise for a wide variety of purposes.
The entire selling point of the CASE methodology, and Oracle Designer, is better programmer productivity and faster overall development efforts. By modeling processes and reusing code, you should be able to shave considerable time off the "time-to-market" of new applications.
We have found that really leveraging the repository data helps us achieve greater gains in these categories. The repository data, as extracted and documented by Publisher/2000, provides:
- Better programmer productivity through automated documentation processes,
- Improved communications between distributed developers or groups, and
- Programmer incentive for using the repository.
With development schedules always pressed, few developers really have the time they need to produce quality documentation of the systems. Yet good system documentation is critical to the development process, as well as to the ongoing maintenance of the systems.
Publisher/2000 creates the documentation automatically and dynamically, almost as a byproduct of using the repository in the first place. This frees up a lot of time that developers would otherwise spend writing documentation and helps us meet tight development deadlines without sacrificing documentation.
At Sequent, we use some of the prepackaged document designs, or rulesets, that come with Publisher/2000. We have also used the authoring tools with Publisher/2000 to design our own rulesets, creating documents that meet our specific needs and format requirements.
For example, an internal life-cycle process at Sequent has its own required documents due at completion of each phase. These project documents are published on our intranet and are used by various project boards for review. We used Solutron's Author/2000 tool to create customized rulesets that generate documents matching these requirements. So far, two projects have gone through the life-cycle process using Publisher/2000. As a result, we now have ruleset definitions created, so that other projects can use this work to automatically create the deliverables they need for each phase of the project.
Other documents that we use frequently include:
- Functional specifications,
- Technical overviews,
- Glossaries, and
- System analysis and table usage documents for program modules.
The Worldwide Data Management Group manages projects that span Sequent's international offices. We used Publisher/2000 on a worldwide human resources project to document the downstream systems as program modules. The resulting documents helped us coordinate activities between the headquarters in Beaverton, Oregon, and offices in the U.K.
As implemented at Sequent, the repository is a large application. Getting everyone to agree to use this methodology and the repository is sometimes an uphill battle; people don't like to change their ways of working, especially when they're under deadlines. Developers need to be convinced of the long-term value of this methodology before they use it.
Publisher/2000 makes using the repository more attractive. Most developers don't want to spend time documenting systems they'd rather be designing or testing something new. Because Publisher/2000 generates high-quality documentation automatically, it provides a strong incentive for developers to use the repository when designing new systems.
This helps us sell the worth of the repository, and the whole CASE methodology, to other developers. In a sense, we can say, "Use this methodology, and you'll get your documentation with no additional effort."
One thing that we've done that has been very useful is to build an enterprise application in Designer that accurately reflects the company. It houses the reusable objects that we share across applications. This is extremely helpful for new application design.
If you're going to go through the effort of creating and managing the repository, it's important to get as much buy-in as you can among the developers to actually use it. It's also important to really leverage the information in the repository for development purposes, for communication with users and for communication between departments and groups.
Sequent Computer Systems
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