Really Strategies
RESOURCE CENTER
“Really Strategies provides us with the third-party expertise we need.”
USATODAY.com
Many Systems Treated as One: Case Study of CQ Press's Content Management Environment
Nov 2002Washington D.C.-based CQ Press is a good example of why a publisher's content management needs are often best met through multiple, integrated systems. In their case, the term content management "system" actually refers to an environment comprising many different processes and systems, not one über system that does it all.
Most of CQ Press's content comes from its line of books. Following composition in Quark XPress, these books are converted to XML and structured in adherence to CQ Press's book DTD. They are then loaded into an application that enables Web editors to manage the content. This system is what most people would refer to as a "content management system," although for CQ Press it is only one piece of the content management puzzle. This CMS for book content stores relational data and XML documents in a Microsoft SQL Server database. A web-based interface allows users to load XML documents into the system, assign metadata, and publish the documents to a number of CQ Press web sites. Metadata can be assigned to any section level in the XML documents, not only to an entire document. Corel's XMetaL is used to edit documents.
"Documents from our books are published to many different web sites, and it would be extremely difficult to manage all these documents and their metadata without a CMS." says Jennifer Ryan, Director of Electronic Development at CQ Press.
Other CQ Press content is managed through different systems and processes. For example, CQ Press maintains a database of bibliographic information as well as a database for images, both of which feed different functions on a number of their web sites. CQ Press recently built a unique relational database with Supreme Court case content for its Supreme Court Collection product. These databases have their own user interfaces for maintaining content.
To manage the content of its weekly publication, The CQ Researcher, CQ Press uses a different process that includes conversion of a Quark file into XML and assignment of metadata through a different interface. "The environment for The CQ Researcher is much different than the environment for processing our book content," says Ryan. "The CQ Researcher comes out each week. The relationship between the print and the electronic is different than it is for book content. The web is a mirror of the print, but if there is an error in the print, we don't go back to press, but we do make the correction on the web site. Moreover, unlike books, there is no ongoing updating. Thus, we use a combined print and web process," she adds.
Why didn't CQ Press build one system to handle all of its content? "It just didn't make sense from both an architectural and monetary standpoint to attempt to combine all of these systems into one. The content that each system manages has different structures and business rules associated with it. Trying to wrap one system around all of that would have been very costly and more trouble than it would have been worth," responds Ryan. In addition, keeping the systems separated allowed each to be developed according to a schedule that made sense for the products dependent on that individual system, rather than creating technical and project dependencies among the systems. Such decoupling of systems will also make it easier to deploy system upgrades, because fewer users will be affected by any one change.
Multiple systems might sound like the right answer, but how well does it work in practice? The answer is determined by whether the systems require integration, and how that integration is accomplished. The need to integrate content management systems is directly related to two factors: workflow and data relationships. There is nothing inherently wrong with having separate interfaces and systems when there is no intersection of the data from a workflow perspective. Even if the same CQ Press user updates the bibliographic data as well as the XML documents, if they are separate and independent tasks in the workflow, two separate systems work just fine.
Dealing with data relationships that cross system boundaries can be tricky. CQ Press has taken the standard approach of allowing some fields to be replicated in multiple systems, but permitting only one system to "own" the relationship; that is, there is only one system through which relationships can be created or modified.
CQ Press staff keep their eyes open for possible areas for improvement in this regard. For example, CQ Press web sites currently combine the voting outcome of a Supreme Court case (from the Supreme Court content management system) with a written summary for the same case (an XML document from the book content management system) for presentation as one web page. For several reasons, the relationship between the Supreme Court database record and the XML document is stored and managed through the Supreme Court system; this requires that book system document IDs are replicated in the Supreme Court system. However, these relationships between XML documents and database records are a growing feature of CQ Press's web products, and to make it easier for web editors to maintain the relationships, CQ Press is now exploring ways to tie its XML content management system user interface into the Supreme Court database, as well as other databases.
Besides databases and interfaces, CQ Press has various processing scripts that help move content, either in and out of systems, to a publishing outlet, or to different content formats. All of these processes need to be considered part of that large-scale definition of a content management system. For example, CQ Press recently created XSLT scripts to transform XML book content to text files with Xpress tag markup for use in the regular print production cycle for the next edition of the book. "This is a new process we are implementing that has been relatively inexpensive when compared with paying an outside compositor to output the old Quark file into XPress tag markup. Additionally, if we made content changes to the book during its life on the web, we can capture those changes for the editorial staff for future editions of the books," says Ryan.
CQ Press's experience demonstrates several lessons for other publishers: First, when planning for content management changes, all the different types of content, the data repositories, and the workflow processes should be considered. Then, the publisher will need to make challenging decisions on how to tie these systems together so they are easy to use and support the publishing channels. For some publishers, this might be one giant system. For others, especially those who, like CQ Press, manage disparate types of content, it is likely to be a number of systems held together by workflow processes and common sense.
