The maxim "Keep It Simple, Stupid" is a useful guiding principle for IT professionals. Unfortunately, it will only get you so far. As IT experts, we're often faced with the problem of explaining complex systems to decision-makers. Depending on their level of sophistication, the result can be depressingly comical. For instance, politicians famously end up describing the Internet as "a series of tubes."
We're just as guilty of oversimplification as politicians. Boosters of Web 2.0 rhapsodize about the transformative effects of technologies such as Ajax and how they will improve the user experience, without considering the other effects they'll have on your systems. In her discussion of the dangers of adopting Web 2.0 technology, Lori MacVittie lifts the veil of simplicity to dig into security issues as well as network and server load concerns. She also mentions that tools once reliably used to measured Web site traffic will likely be broken by the use of Ajax and its ilk.
Here's another example: Many vendors tout e-mail encryption as a simple solution to a range of regulatory mandates around information protection. But before you buy into that notion, read Christopher Beers' special report, in which he discusses various approaches to implementing this technology. And while Christopher does an excellent job of laying out the technical considerations, Patrick Mueller's column points out some of the legal ramifications (dare we say, complications) of e-mail encryption. Considering one without the other will most assuredly not get you the result you want.
One might guess by these examples that I'm encouraging a lengthy planning and review process designed to consider in minutia all the nuances of technical decisions. I'm not. Careful planning is certainly the hallmark of a successful IT organization, but you can overdo it. Analysis paralysis is all too common in our industry. What's called for are carefully devised real-world tests, done with the expectation that the systems you test will be flawed and that you will learn from those flaws. It's a methodology we hold dear at Network Computing. We know from experience that you can analyze any technology--or for that matter any complex decision--to any degree you like, and you'll still be surprised when implementation issues arise. Analysis requires assumptions, and assumptions can be wrong, even if only subtly. Testing verifies assumptions, so don't leave it out of your process.