I've made a similar analogy in another thread to the ones below, but it bears repeating in the context of this post. The best analogies come out of the automotive industry, that I've been exposed to all my life (over 40 years) as the son of a dealership manager.
1. A New client built on the platform of an existing one does not mean that the old one was a failure (though KR was fairly crippled to start out with).
After all, every year, every car manufacturer puts out new model year vehicles, about 80-90% built on the same chassis as the previous year - just with upgraded parts to better fit who they think are the targeted customers, or to fix bugs reported by QA or the current owners. Half the rest are "new" models, with new names, but built on the existing chassis of current models (examples: the evolving of the 90s Monte Carlo from the Lumina, which in turn was based on a stretched Corsica/Beretta, etc. And, the H2 & H3 Hummers, that used modified versions of existing GMC/Chevy SUV chassis when developed).
2. Sometimes "just repairing" isn't an option. You see this all the time on TV shows about restoring classic vehicles. Sometimes there's enough left of a car to restore, but you just can't bondo and paint it, and expect it to last.
When these people "restore" a car, what they mean is they strip it down to its component parts, refurbish (not just repair) what is still structurally sound, and replace (sometimes having to hand-craft) what isn't salvagable or was just plain missing when they found it as a rusting heap. They then effectively rebuild the car from the ground uo, with all those restored & new pieces, some of which are often far better than what the car's original equipment was, if not going for a classic competition competitor vehicle (where cars are judged by the parts numbers, and often even the serial numbers, matching up exactly to the OEM equipment for that manufacturer, in that year, for that specific model).
Since that's not the intent for a rebuild for improvement, they'll definitely be adding in new pieces that weren't present in the original KR, or even the version patched up to today. But replacing items patched several times, with items that do what the patches do out of the box, will vastly improve performance out of the box, for ANY software.
KR's flaws, IMO, can probably be traced to EA Japan. Unlike EA here in the states, they promised KR as a retail box product by a certain date. So, in order to not lose face (And the economic repercussions of not meeting that deadline, in terms of manufacture & delivery), they effectively forced KR out prematurely. The Beta test was with software that was at best ready for Alpha testing, and despite all the open beta testers (and probably the programmers) telling them that at best, the software was in a state that it should have been for in closed beta for a few more months, it was shoved out to the world market (even the places it was a free download) to meet that impossible deadline. Just my feelings, based on what I know from having to deal with a similarly implacable Japanese industry (anime) for 10 years as an anime convention staffer (and sometimes department head), getting to listen in on horror stories (often over dinner) between American/Canadian industry people who've worked on projects over there, or dealt with deadlines in the US imposed by the Japanese owners. Western companies are much more likely to admit something's wrong, and push back the release date of something to attempt a fix. Japanese companies (and the Japanese divisions of western companies) are masters at SS Titanic deckchair rearranging, and will go out of their way to cover up issues that are potentially fatal to their product unless the release date is pushed back, rather than admit something is wrong, and ask for more time (remember the issues over the last few years, with various game consoles being released with severe issues? - same issue, different expression).
