On countless occassions throughout my career, I have been tasked with fixing up legacy business-critical applications containing old, fragile code, that may or may not have been designed very well the first time. My initial reaction is usually "Just rewrite it," but that has never (and I mean never) been the desire of the business. Often, the business has invested so much time and money into their product that there is an institutional pride which precludes even bringing the subject up. I'm sure that this is not an uncommon situation for many developers out there.
Essentially, we are starting with a brownfield project, and trying to make it act like a greenfield project. I call this fertilization. I will be writing about this process in detail in the weeks (possibly months) to come.