Quoin worked on a series of projects for the WashingtonPost.com:

Web Publishing System – CMS Delivery Engine

Quoin built a content delivery system for WashingtonPost.com. The CMS Delivery Engine delivered by our team supports one of the most frequently visited news sites on the Internet, and delivers more than 200 million page-views per month. As a national media outlet for political news and content, the client had critical requirements for scalability, extensibility, and security of its online publishing systems. This system uses a robust J2EE architecture for delivering static, dynamic, and personalized content. The initial phases required the migration of a mission-critical system that was built on a proprietary template engine to a standards-based solution. The new architecture has proven remarkably extensible - the team added significant functionality to support the personalization features required by MyWashingtonPost.com in just two weeks. The figure below shows the system architecture for the new publishing system and integration with the editorial system.

CMS Delivery Engine System Architecture

The delivery system was responsible for generating all static XHTML pages or rendering dynamic page components. Our team implemented a three-tier architecture comprised of: Java Server Pages for presentation; Stateless Session Beans to implement publishing processes or rules; Domain Objects implemented in Java to manage persistent data. We also implemented a production database in Oracle and the publishing process to transfer content from the legacy editorial system. (The reengineered editorial system shown in this figure was planned, but never implemented. WashingintonPost.com replaced both delivery and publishing systems in 2005.)

Application Generator for Reader-Searchable Databases

Quoin built an application generator for use by WashingtonPost.com editorial and production staff. The toolkit is used to specify and generate functional J2EE applications for reader searches of online databases that could be quickly deployed to the CMS Delivery Engine. The project objective was to substantially reduce the time-to-market for database search applications. The client, constrained by developer resources, typically required several months to deploy a new reader service. After Quoin delivered the toolkit on-time and at its fixed cost, the client has used it to successfully build reader-searchable databases within a three-week development cycle - yielding more than a 4:1 reduction in development schedules. The editorial staff has used the toolkit to create databases on elections, schools, auto reviews, and other topics. These applications have proven remarkably successful at increasing readerships and advertising revenue. Shown in the figure below is the workflow for the application generator.

Database Application Generator Workflow

A user will create or update a project using the application generator. This process will generally require several days to preview the schema for the database, define the summary queries for high-level results, create the detailed queries for drill-down by readers, and map the outputs to keys used in the schema. Although the toolkit requires a basic understanding of relational design, non-technical users have proven adept at rapidly creating a searchable database application.

Results

• Quoin defined an innovative design based on a common framework for reader-searchable databases.
• The design avoided the need to generate application-specific code and the subsequent maintenance effort to modify these applications.
• System enables editorial staff to create applications, and requires minimal effort by the production staff for deployment.
• Database Application Generator has resulted in more than a 4:1 reduction in development time and effort.

Content Management System Requirements Analysis

In 2002, WashingtonPost.com initiated a process to define the requirements and an implementation plan for a new editorial system. Quoin was engaged to lead the analysis, and to produce a validated functional and business requirements specification. Our project team worked with the editorial, business, and technology staff to identify these requirements. We produced a use case model, including detailed user stories and functional requirements. In addition, Quoin conducted an analysis of the current business processes related to content management, and defined a possible future state to improve business and technology alignment for the organization. The results of this project were key resources when WashingtonPost.com selected a commercial content management system in 2004.

Results

• Quoin defined a validated functional and business requirements specification.
• We engaged editorial, business, and technology stakeholders in requirements analysis.
• Requirements analysis enabled the client to select a commercial CMS.

Content Management System Performance Tuning

WashingtonPost.com implemented a new content management system in 2005. However, the editorial system exhibited a number of performance problems, including slow response for many key functions and limited scalability for the required number of users. Quoin was engaged to help diagnose and resolve these performance problems. As the principal consultant for this effort Eric Meyer collaborated with the technology staff to define a pragmatic approach for quickly improving performance of this critical system. He defined a strategy for benchmarks and metrics from the editorial system, WebLogic Content Server, and Oracle content repository. Eric also identified a number of potential causes, including the non-normalized database schema, custom object-relational mapping components, unnecessary remote method invocation, overhead container-managed Entity Beans, and other aspects of the system. WashingtonPost.com staff implemented his approach, and Eric provided continuing guidance. His performance analysis strategy was key to the eventual resolution of these performance problems.

Results

• Quoin defined a strategy for analysis of significant performance and scalability problems with the editorial system.
• Our guidance enabled the WashingtonPost.com staff to diagnose and fix the problems.