Quote Originally Posted by Talraen View Post
Again, I don't think "storing" data is the issue. But your point is well taken, there are definitely issues that affect future design. That said, I don't see how the game will become stale simply because character inventory is permanently locked at 140 slots. They can add retainers, chocobo saddlebags, and who knows what else ad infinitum without worrying about the issues that prevent inventory from growing. They will have an issue if we get past 35 classes/jobs, since the armoury is also linked to this issue and will need to be expanded again. That does give them a few years to address the problem. Assuming the game remains popular long enough.

What specifically are you worried about becoming "stale" due to these issues?
In my job, I see a lot of people make decisions based on the problem in front of them, and not the root cause. In this case, you've mentioned the frequency at which data is sent being the prime issue. That is a problem for sure, but if the data were stored more efficiently, it suddenly becomes less/not a problem. It's possible there's even a deeper root cause than what I can see based on my limited understanding of their code base, but I'm VERY confident that the issue is definitely deeper than merely the transfer of said data (which is a problem).

Things like menu-based content, opportunity costs for implementing band-aids that could go to better fixes/new content, gearing/itemization depth, etc.

Quote Originally Posted by Vhailor View Post
It's not a question of whether or not data gets sent; it's how much, and how often. FFXIV sends huge amounts of data very frequently. The net result of this is less third-party manipulation (in theory - there are plenty of exploits that FFXIV has had to deal with), but a lot of unnecessary constraints imposed by server limitations.

There isn't a perfect solution, to be sure, but I would argue that as soon as an implementation begins meaningfully infringing on a company's ability to grow the game, it's time to shift course. If SE doesn't want to rebuild the engine, they should invest in vastly improved server capabilities, so that the effects are invisible for the foreseeable future. Otherwise, they're letting their own architecture throttle their cash cow.
Bingo.

Quote Originally Posted by kikix12 View Post
The server doesn't care about how you look. Graphics are applied on clients side. Your character could hop from server to server and none of them should ever send you data on how your character looks, because it should be stored on your clients side and shown to you from your clients data.
I don't know the specifics, but what I do know that is SE utilizes client-side operations the least out of any of the MMO's I've played over the course of nearly 20 years. In fact, IIRC nearly everything is handled server-side via checks. This has benefits (mostly security/anti-cheating measures), but it also has numerous consequences (performance, stability, etc.). The reason that so few companies employ this methodology is because it's NOT the most efficient design. Typically the drawbacks don't offset the gains, and IMO FF14 is a good example of this being true.

Quote Originally Posted by Tridus View Post
Yes, it's this. The root of the issue is that the game resyncs your entire inventory between client and server far, far too frequently. Effectively, it winds up sending your entire inventory because you looted something, instead of saying "you looted something" and letting the client handle that event. While that does nicely prevent the state from getting out of sync, it's horrifically resource intensive and that's why other MMOs don't do it.

It's also why they implemented things like Chocobo storage instead of simply adding more inventory. Chocobo storage is only synced when it's actually open, and it isn't available at all in many circumstances. It's a means of working around the root problem.
I vaguely recall reading one of the tech articles from SE going over how the server validates an operation. It was the most asinine thing I ever read. You push your hotkey, which sends the server your "intent" and the server sends an OK back (if OK), and then your system resends the press and the server validates that what you sent matches your intent.

As far as their band-aid solutions I just wish they would devote that dev resources to an actual solution. I understand why they don't (cost), but like I've echo'd before. It's likely going to be a problem eventually and the longer they wait to fix it, the harder it will be.