I think I'm erring on the side of sadness. In writing this - the subsequent conversation I've had with bambuka and others who know the code far better than me, the conclusion that I've come to is that the existing codebase is unworkable. (I won't repeat what I said to bambuka below (or above depending upon what order the comments are 🙂))

I think that any developers who already work with or try to work on steemit.com in the future will experience the same issues that I've had and will almost certainly come to the same conclusion. The level of specialism required is hard to come by, which means it's expensive. In the long run, Steemit will be much better off starting again (not completely, but I mention that to bambuka 🙂)