I'm curious - do you think FFXIV would really become a transformed experience, one well worth its monthly sub, if technically-limited QoL features like the Glamour system and housing and character creation system were completely resolved to everyone's satisfaction?
To me, the core issue isn't a lack of QoL features, it's boredom. Certainly I'd love to have a Miqo'te with a proper ass again, but at the end of the day, it wouldn't entertain me more than I'm entertained now. Shallow gameplay surrounded by overhauled and improved customization and UI systems is still shallow, to me. I think these problems are easy targets for complaints, but to me, it's a very surface-level issue that SE would be wise to back-burner in favour of creating enduring, longer-lasting content.
Oh, man, yes, the code is awful.
Firstly, due to server sharding, there's really no difference between 600,000 and 6,000,000. Just add more servers. And as an aside, we're nowhere near five million. Not even close. That would be around 75,000 people per server on each individual server. That's too large by an order of magnitude. Do you remember when SE was splitting map instances near the launch of Stormblood to handle the temporary influx of players? Yeah. They used three instances, and none had more than 150 people in it at any given time. I'd be surprised if there were more than 5,000 truly active players on any given server, and the true number could easily be less than half that. Nowhere near five million total players.
Secondly, the engine was awful at the start. It's based on FFXIV 1.0, which, for its time, looked like utter ass considering the resources it required. Sure, character textures were decent and polygon counts were high, but textures and level design was noticeably repetitive (literally there were copy-pasted chunks of map), and there were plenty of single-player titles that looked as good or better while drawing less power. ARR solved this to some extent by bludgeoning the engine's technical side (texture resolutions, poly counts, etc.) and attempting to gloss over it instead with vibrant artwork, but that approach is rapidly showing its age.
What worsens this, of course, is that SE routinely makes the bone-headed mistake of designing content that squarely collides with the limitations of their crappy engine. Eureka is a prime example of this. The idea of an instanced, cross-data-center environment where players can team up in large packs and have a different experience sounds good on the surface. Until you remember that we can't send /tells from instances, obviously can't join a friend's party if they're in the instance and we're not, can't Black List from within instances, can't form truly durable Linkshells with the people we might meet in instances, couldn't at the time consistently target party members or monsters when enough players came together in one spot... these issues should have sent Eureka back to the drawing board. Instead, SE stupidly plunged forward, and released a piece of content that will in all likelihood never fully realize its potential. Or, look at Housing: SE devised a tremendously ambitious Housing system, with nary a thought paid toward the fact that their existing server infrastructure was wholly inadequate to address the demand. The result was players waiting literally years for the Housing supply to even remotely approach demand, and the wait is ongoing on more crowded servers.
Look, I get the desire to defend SE sometimes. I can certainly be harsh on them myself, and so can the forums. But FFXIV's underlying code is not a hill upon which to make an effective stand.