Actually, the Right Way to work on a network is to make absolutely sure you have a fallback plan so it's not dead on Monday. In my experience, and I've worked in IT departments, the reason the network is down on Monday is people thinking "this is going to be easy, we'll just blow it out there and BAM, we are done". Somewhat like pumping a bunch of text and picture links into an HTML file, checking it on a single browser (but not against a compliance validator) and calling it good.
What I was trying to do was:
a) Use new features, such as MathML b) Look to the future to keep upgrade needs (and non-compliance complaints, ironically) to a minimum
In order to qualify as the Right Way, a method must work. The reverse is not true.