The challenge of integrating software and systems will always be with us. In the brief but turbulent history of information technology, creation and destruction go hand in hand. Old technologies and approaches give way to new ones, sometimes quietly and sometimes with a fight. Yet, in this maelstrom of activity one thing remains unchanged. Our desire to solve bigger and more important business problems breeds increasing complexity. To battle this complexity we divide and conquer. We don't want to reinvent the wheel. We try to wrestle control of problems that we know how to solve and then we build up from there to a final solution. The larger pieces become independent systems and applications. Integration - broadly speaking, the superset of EAI, B2B, BPM, and many other acronyms - is the process through which the larger pieces work together to solve problems.

In my article in this issue,"Integration Is the Killer App," I argue exactly what the title claims. Integration is how we extract value out of systems above and beyond what they were originally designed for. In a connected world, where more and more applications and data sources are available both within and beyond the firewall, integration is the key to managing information and leveraging network effects.

Integration would be easy if only the problems we wanted to solve had many similarities and if the methodologies and tools for decomposition and recomposition were well developed. Of course, this is not the case. In Eric Newcomer's words "the software industry remains essentially a craft business." The problems we are trying to solve are very diverse, even within vertical domains. We reinvent the wheel every day, multiple times. We cannot agree on methodology, perhaps because we have not developed a good one yet. Well-trained integration personnel are scarce, even in this economy. Therefore, we have to put our hope for making integration easier and more efficient in the development of integration technologies and best practices.

XML and Web services are our current best hope for lowering costs across many of the facets of integration.

This month, XML-J addresses a number of important areas for integration. In "Can Software Cross the Standards Divide?", Eric Newcomer looks at the relationship between standards, competition, and innovation. Ayesha Malik explores XBRL, an emerging standard for expressing business information, in her new monthly column addressing standards activity. While it may be easy to standardize on the basic mechanisms for exchanging messages between applications, it is not always easy to have meaningful interactions. One way to add meaning is to agree on vocabularies. Kristian Cibulski's article, "An Introduction to BSML," introduces the Bioinformatic Sequence Markup Language (BSML), a standard mechanism to communicate genomics research information. When two applications cannot speak the same language, translation becomes necessary. XSLT can transform XML documents from one vocabulary to another. Yuhang Sun's article, "Developing Complex XSLT Scripts," shows some useful patterns for building flexible and reusable stylesheets. One way to speed up integration is to efficiently access and manage data in XML format. However, no matter how easy it is to get access to data and to move it between applications, there is always the business logic to worry about. While it remains buried inside custom applications, the cost of implementing changes to integration projects will be high. The solution is to externalize as much of the business logic as possible through workflow specifications and business rules. The move toward standardizing Web services choreography is a step in the right direction as Suhayl Masud shows in his article, "Building a Real-World Web Service," which applies BPEL to RosettaNet processes.

This issue takes a balanced look at the progress we're making and the direction we need to go to solve integration problems, along with some of the obstacles we're bound to encounter along the way.