Conversations this week with folks in multiple, different organizations I have connections to, about formal change controls. The general rule, especially for the non-public companies, seems to be reasonable levels of process, but not as well documented, automated, and not as rigidly enforced as the more rigorous among us would prefer.
It was interesting talking with the groups that were most frustrated; the developers will talk the talk, but find ways to get around the process when crunch time comes up. Of course, some will use the “boss needs this ASAP” line, but when I check with those who know their management, there is really no desire to circumvent tight process just because someone says so.
The challenge seems to be where to begin, and making the requisite time to “do it right”. How often have we made policy, captured it in an eMail and sent it ’round to the usual cc: list … and then are surprised a few months later when the same conversations come up anew.
My favorite way to tackle any documentation or process design that is at the lower end of the priority scale is what I call my Ovid approach (AKA). Each time you touch the process, just improve the documentation / design / structure a little more. Before long, you’ll notice a really nice document!
Automation is another powerful process enforcer, a favorite in our shop. It’s a lot easier to control when stuff gets pushed into production when there is only one way, automated, audited, and reversible. For us, when something needs to get on a web site, say, by 8am the next morning, that really means it’s gotta get to the staging server by 2am – else the automated push will have to wait until the next night. A structured, documented, auditable process is quite easy to “blame” when folks grouse about the lack of immediacy – quality first, yada yada.
Mind you, this is actually a hard push-back on the folks making these complaints / observations. It’s generally easier to be frustrated with the nature of things than it is to do the proper documentation / automation.