As I said it is a major release type of project as far as I'm concerned. If SE had stepped up to the problem a release or two ago we might have been able to hear that it will be fixed in 5.0. I still think it is less to do about money and more to do with dogma and saving face.
I had over 30 years of software development before I retired and still consult. I started my career in mainframe and Unix operating system I/O subsystems (i.e. device support) in the late 70's. I was a company representative to the IETF during the 90's and oversaw multiple areas of our TCP/IP suite architecture and design. I'm also well versed in other communication protocols. I've architected and designed large relational database middleware products still in use in today.
I am well aware of the potential problems of implementing new function into existing code. It is one of the reasons I think any new housing system should be done in parallel to the current ward system. It is also why I think they need to create self contained instanced housing system. The odds of breaking anything are much less when starting off with new code and designing it with an extensible structure in mind. I can determine this without having to know the intimate details of the current ward system. Based on SE's comments I can also make a pretty good guess that the current system is a static one as far as system resources. Developing a dynamic platform for instanced housing would increase flexibility, reduce system management and reduce overall system resources when compared to the current housing system. I also feel, without looking at the code, that a good portion of the actual application code for the dwelling, exterior and client could be reused saving on time and costs. Like any project it will require people who have a deep understanding of the existing product and those that have good understanding of systems and databases to do it right. It would also behoove them to have someone with a good understanding of how the various programming languages work. One thing I've heard more than once from SE is part of their server capacity problems is the duplication of large data areas passed between game functions. Whether it affects the current housing system I don't know but these types of problems can be resolved by programming language selection, changes to the way data is passed, etc. However it does cost money to hire people with this knowledge. Will SE hire the skills they need? Who knows but their history suggests probably not.