Sorry as a dev I'm going to disagree with you here.
Depending on how much changes you want to make it is definitely easier to Redo a project from scratch then to re-factor.
Fix something here, causes other things to break, then fixing the things that break causes other side effect. Chasing some unknown bug in code you did not write yourself. And then spending hours and house on trying to think what went on in the other developer's mind when he did this or that, wastes a lot of time.
I have a few years ago agreed to do continue and debug an existing system (Granted it was not written with software I was familiar with at the time). It was a nightmare. If I had re-written the software (Even in the same technology) it would have taken me 3 months of lost time, but bug-fixed thereafter would have been hours instead of days.
Instead I continued on that broken piece of software with infinite bugs for about 5 months until I decided to pass that problem on to someone younger than me, with more energy and optimism and more desperate for money then myself. The owner refused for the application to be re-written as they already paid for the development (Sunken cost)
Seen as this is a project that DC still want to work on for years to come, I can see his rationale for re-writing it.
I kinda agree -- I was on my way out and I only mentioned the hassle of refactoring into a spiderweb of code.
You bring up a good with the trades made -- and DCs not completely affected by industry norms, which is why he chose the latter -- longer, but better way. The norm, I mention is that rewriting from scratch (or similarly porting systems into a newer/updated framework) is not an option for companies who have their apps live or have users already relying on them.
Turning off a service for 3 months just for a rewrite will generally piss off customers and have them leave for an alternative. (This from personal experience, is why a company I worked for refused upgrading their systems from JS ES4 to ES7+ so we could use some of the QoL syntax in newer versions)
Since DC is in a unique situation in that SS is not a live service -- he still has a consumer base to think about.
The reason Debbie is such a meme at this point is that she was the first piece of content in 2016 -- and she has been reworked several times prior to the "tech-update" and now she is being redone yet again.
The reason why I don't think is valuable use of time to rewrite and start from the beginning like this -- using Debbie as an example -- is that she had a few janky animations (bed scene, shower scene) pre tech update. Her animations were "done", her story was "done" and in many ways -- none of her code should affect how any other girl functioned.
So lets say they started using a new rigging system for the newer girls (pre-tech update) that new method used a different framework than how old Debbie was being animated. However using that new method would not affect Debbie, and Debbie's code did not affect Roxy or Consuela. Instead of moving forward -- they went back to Debbie and Jenny to set them up for the new framework. This eventually let them delete the old code since everything would be running on the new framework.
I think this is the best practice -- but it has led to the Debbie meme. And this was all pre-tech update stuff. To do that again starts feeling disingenuous. How they handled updating Debbie the first time around where Debbie's garbage code was eventually* replaced with the newer framework utilized by newer content -- is what I meant about the value of refactoring. These new updates utilizing whatever new framework and starting from scratch is suspicious. I mean, I do see the difference. I do appreciate the differences. There are minor rewrites even in the new updates that are quite noticeable.
I still commend them for all of that. But its still suspicious what actually has been achieved in the last 2-3 years. There was 2 years of silence, of no updates, and you could assume they did what you alluded to by taking off "3 months" to start from scratch and move forward after -- because it was suppose to be faster. But now after 2+ years for the framework rewrite (and then seeing a sliver of code from pre-tech compared to a recent build -- with nothing notably too different) and lastly the idea of "suppose to be faster" -- yet these Debbie updates have taken 6 months -- continues to raise further and further suspicions.
I don't think my point was ever really about refactoring or rewriting code from scratch, tbh, I think I got triggered when what people say sounds like they are blindly defending DC. And a terrible code base is not a good enough excuse for what has happened in the last 3 years. And I just wanted to qualify that, hopefully.