Knee-jerk Architecture

From the Spoutlet, one of the most entertaining (and unfortunately accurate) IT blog posts of the year...

My man Jimmy's been <a href="http://www.jnsk.se/weblog/posts/whatprocess.htm">bitten by the process bug</a>, but it doesn't seem to be anything major. Although I don't consider myself a methodologist, I thought I'd relate a recent story on the topic.
<BR/><BR/>
I'm spending part of my time these days as chief architect on the development of a mission planning system. Anyway, this project was running into some budgetary constraints for the current fiscal year. The problem was that the impact would be felt like aftershocks in the coming years. There were three critical modules that were hit by these constraints.
<BR/><BR/>
When asked to describe how we were going to develop the system, I began talking in an "implement-by-slice" language and putting dates beside each slice, finally showing what could be completed by the target date (quite a bit less than many stakeholders had in mind). One of the dev leads that was with me wrote it up, and sent it out with the appropriate explanations and commentary.
<BR/>
The next day I was busy in meetings all day long with our local Microsoft branch, who are absolutely amazing, by the way. The following morning I got a call from the project lead: "We've got a green light for all the modules - budget, staff, everything." I didn't connect the dots, I mean, I just wrote up a fairly standard development plan, no reason for that to have any serious effect, right?
<BR/><BR/>
Well, it turns out that this bit of process, implement-by-slice, was new to many of the stakeholders, and when they saw that the system would be stable at numerous points of the project lifecycle, it blew them away. This down-to-earth plan became a framework for concrete discussions about priorities and budgets. It forced everyone to stop daydreaming about what would, someday, be possible, and focus on what needed to be done at every step of the way.
<BR/><BR/>
--
<BR/><BR/>
Anyway, this isn't a commercial for implement-by-slice. Rather, consider it a way to regain focus; while many of us talk about how process doesn't matter much, I think that that may not be so true. While it may be well backed in terms of the development team, in the larger context of the entire project, process - in terms of the kinds of documents that get circulated, can make a big difference. Lets not fall into the <a href="http://www.pyrasun.com/mike/mt/archives/2005/07/13/21.45.29/index.html">"knee-jerk" trap</a> (via <a href="http://pluralsight.com/blogs/dbox/archive/2005/07/14/13346.aspx">Don Box</a>).<BR/>