By Jeff Gainer
(Author’s note: This article appeared in the June/July 1998 edition of the itmWeb Report.
In the last two installments of Critical Path, I presented overviews of two approaches to software process improvement. I'd like to conclude this discussion with a few words about the key to both: documentation. Tom DeMarco is often quoted as saying "You can't control what you can't measure." Similarly, I would like to postulate that you can't repeat what you don't document. The Capability Maturity Model is a veritable recipe for software process improvement, yet many of the companies I consult to regard it as an impossibly complex enigma. But there is nothing complex about the model, much of the work in implementing process improvement is the difficult first step of documenting processes and procedures. This simple realization was best articulated by a client who said "You know, we already do most of the Level Two and Three practices, the key difference is that the CMM requires that we document what we already do!"
Exactly! And documentation is the key to making the process repeatable! Other than lack of available time, there is no reason why that a well-intentioned software development shop can't do their own informal process improvement using the CMM as a guide. The ever-difficult first step is documentation, followed by implementation, measuring and monitoring.
Report from Boston: The Cutter Consortium '98
In April, I attended the second annual Cutter Consortium Summit in Boston. The Summit is a unique opportunity for IT professionals to hear presentations and debates by industry notables like Ed Yourdon, Roger Pressman, Steve Andriole, and Larry Constantine. By design, the Summit is a small gathering, and thus gives attendees direct access to the speakers. The debate at the Summit is strong, sometimes loud, and often becomes . . . well, spirited. I recall a prestigious panelist screeching "Bullshit! Bullshit!" at one of his colleagues during a particularly passionate panel discussion. I'd suppose that the erudite panelist would rather not be named, but I do admire and commend his eloquence. The following day, Ed Yourdon reminded the assemblage that members of the mainstream press would be joining us, and thus we should tone things down a bit. I don't recall that anyone complied.
A special appeal of the Summit is that there are no vendors; instead, attendees can spend their time discussing issues with the panelists and speakers on a one-to-one basis. While at the Summit I had the opportunity to personally thank Rob Thomsett for helping me to stop feeling guilty about preferring Fortune to Visual Basic Programmer's Journal, even though I am a contributor to the latter. Rob's newest article, "The Care and Feeding of Project Managers" outlines the key difference between project managers and technical managers. It was originally published in American Programmer (now the Cutter IT Journal); the article appears at the Thomsett Company Web site as "Into the Twilight Zone." For those not familiar with Rob's work in the project management field, or his shy, demure manner, a visit to the "Hot Articles" section of the Thomsett site is in order. There you will find his now-classic "Double Dummy Spit," which explains the complex negotiating dance of estimating games, as well as some wonderful bits of Australian slang. I look forward to Rob's future work, particularly that dealing with the care, feeding, and merciless destruction of "bunnies" who don't document their organizational processes.
Copyright 1998, by Jeff Gainer