I started my information systems career in the early 90's as an officer in the United States Air Force, working in a software development group inside a large, secure intelligence center. The amount of data that travelled on that network was all but unfathomable. Textual analysts in particular were struggling against massive volumes of information in a multitude of databases and formats, with news feeds and new systems adding to their burdens each day. They then encountered more adversity getting the results of their work back out to those that could benefit. Large projects were handled by contractors, but the handful of military and government employees in my office were still overwhelmed in the middle of this chaos. The older mainframe world was giving way to the newly dominant client-server model, and not every programmer made the transition well. There was increasing pressure to adopt new methodologies, processes, and languages. As a result, the fresh Lieutenants like myself were given lead roles in some of the most important efforts.

The big catalyst arrived around the time I came on board: the world wide web. Back then, Java was a glimmer in Sun's eye, NCSA Mosaic was the only graphical web browser, WAIS was a leading edge indexing mechanism, and Yahoo was just growing out of its initial manifestation as a text file. However, web protocols and open source products proved reliable and cheap to implement, even in light of defense intelligence security requirements. What began as grassroots efforts to organize information behind web servers rapidly became the strategic direction of the entire community, with lowly coders briefing generals on a regular basis.

Unfortunately, this was happening simultaneously everywhere from the bottom up. With more brass hefting their weight around, the energy became fragmented. The results lacked unified vision, boasting more stovepipes than 19th century London. The situation was summarized in a recent article on government IT:

The Defense Department began mapping an architecture for its intelligence-communications network in 1995, when it realized that it operated more than 300 incompatible intelligence-communications systems. "We didn't know how to compare any two of them," says Truman Parmale, a Defense Department program manager and systems analyst. "They had no common foundation."

The article was about architecture, and I replied with my own testimony.

That same year, at a military software boondoggle in Salt Lake City, I had the opportunity to attend a lecture by software giant Grady Booch on the state of the art. He stressed the inherent difficulty and epidemic failure of large software development efforts. While he saw good news in emerging methodologies, processes, languages, and tools (maintaining that object-orientation was the major mark of progress), his central contention was that the most successful projects were architecture-driven. I saw the light, and returned to the front lines with the concept of system architecture ever in the subtext of my dialogue and development work.

I can't say I witnessed much improvement during my stint with the military, but the article I mention above shows that the government at large has taken up the same calling. It looks like Grady is on track too, as his company Rational is now being bought by the judicious businessmen at IBM. What I will say with certainty is that my own current direction is entirely architecture driven. The central software offering from Mentata Systems, the LDAPHttp Framework, represents a defined platform for delivering systems with well-partitioned business, control, and presentation logic. It was designed with no particular application in mind, but rather a consistent focus on leveraging the best parts of the disparate pieces that comprise the whole. In the end, it will be my architecture of choice for delivering web services.

So with the advent of 2003 I have made my resolution to stop grumbling about what I don't like and start singing about what I do.

Why LDAPHttp?

Happy New Year!
   The hand is better, thanks.