You're lying to yourself if you think that's even remotely true for some of the examples of complaints you brought up.
Printable View
You know, I watched a Mr. Happy video today about Criterion dungeons and he put out a point that makes a lot of sense:
Mount Rokkon's poor reward structure is the same as the previous V&C because it already 'sailed' (production wise) when the first one was out. He mentioned that the developers explained about a similar situation in Eureka, that only saw major positive changes in Pyros, because Pagos was already ahead at a development stage and they couldn't improve by then.
That's probably a symptom of how formulaic they work... The features across the expansion are consistent in what they deliver, so to change something on the go might not be as simple, sadly... But hey, I'm not a game developer, so that might not even be the case.
as we all are of yours ^^
personally, if I was a Dev, why would I, as a japanese speaker come to the English forums? I wouldnt.
thats why if we are going to place blame anywhere, we should blame the community reps, because THATS THEIR ACTUAL JOB!
if the english forums have no community reps, thats on SE to fix. its not, and should never be on the developers themselves to visit the forums. I doubt they have the bandwidth...
They listen to us, sometimes.
The problem is they listen to other people a lot harder, and the people who they listen to want completely contradictory things from what we want.
This isn't quite true...in mentor roulette even though it shows tanks as needed, I queue in and have to wait a bit sometimes. If I queue in as a healer, most of the time I get insta-queued. I think there is some priority system for checking tanks first. As mentor roulette has no limit on what you can be queued in for it shows the true "in need" I'd say...In reality healers are in the need more but the tank privilege is shinning through because that is the class that say it is always in need, despite it not really being so. I have also seen DPS show up as "in need" but very rarely for mentor roulette. Healer though? In mentor roulette I haven't seen it appear..which is odd. Or if it did, it was gone at first glance.Quote:
Case in point with healers - I readily find I have instant queues with tanks, but ~1 minute queues w/ healers. What does this tell us? More people are queuing as healers than as tanks. So, more than enough people are either straight-up enjoying healing, or at least find it sufficiently good enough that they're actively choosing to play it. If anything, it suggests tanks might warrant some kind of tweaks to make them as appealing as healers are now. Again, however we much we might desperately want the world to agree with us, the reality is that compared to the size of the playerbase, you need a truly enormous number of people to actually represent anything significant. The devs listen - and see that healers are okay right now.
devs listen so well that they completely ignored all feedback related to Kaiten that they specifically asked for to the point of just skipping over even mentioning samurai in any pve related live letter since (afaik anyway, i stopped paying attention to them after 6.2 came and went).
It can go either way, and there is issue where the updates to roles needed are too infrequent and sometimes straight up delayed. The priority list is overall Tank > Healer > DPS, but because it updates with delay it often is inaccurate representation of actual needs. I had it go both ways, queueing in healer needed as healer for instant pops, and queuing for tank needed for instant pop as healer. The only thing thats actually clear from that is that tank and healer are, overall, less played than DPS but thats true of p much any mmo.
Let’s ignore that every metric of job satisfaction and level of play relative to the other jobs crashed when they changed healers in ShB and even with a new job this expansion that was billed as being build for modern healing rather than lobotomising an old job healers still are overall the least played job and most often the ones with the shortest queues
Yeah but that doesn't explain why in mentor roulette healer role needed is nonexistent where I have seen dps needed is what I'm saying. I think it's tanks>dps>healers where dps frequent too much of a majority when queuing for normal things that it offsets and skips straight to healer with the exception of alliance raids as a majority needed are DPS which would make sense given more likely dps needed is seen on them {I imagine this might be due to party queuings where healers and tanks are accounted for in the queue so it takes DPS even more so}Quote:
It can go either way, and there is issue where the updates to roles needed are too infrequent and sometimes straight up delayed. The priority list is overall Tank > Healer > DPS, but because it updates with delay it often is inaccurate representation of actual needs. I had it go both ways, queueing in healer needed as healer for instant pops, and queuing for tank needed for instant pop as healer. The only thing thats actually clear from that is that tank and healer are, overall, less played than DPS but thats true of p much any mmo.
That's because the ones queuing are dps so I imagine it skips to healer in priority order because dps are the majority of players as mentioned by you. Since lots of dps are queuing for it, it goes next in line outside alliance raids and trials sometimes where the dps may be seen needed because it requires more of them. It makes sense logically if that is the case as to why I've never seen healers appear in mentor roulette needed, despite them being an insta-queue and tanks sometimes not.
The adventurer in need doesn’t reflect actual queue times because the system is designed to check tanks first
Say in the last 10 minutes
-50 tanks
-20 healers
-300 DPS
Have queued for the expert roulette, the system has a set amount based on queues that it will use to determine what constitutes being adventurer in need, say it decides that if there is less than 60 of a role that role is adventurer in need, it will check tanks first, see there less than 60 tanks and list that as adventurer in need despite there being less healers overall whos queues end up popping faster
It’s also why at 3am server time when all the queues are dead the queues all basically default to tank being adventurer in need; because there isn’t enough people for any of the roles but tank is checked first
About the only thing you can use adventurer in need to determine is if there is DPS in need for any roulette other than the alliance roulette then there is a current shortage of DPS at that singular time point, adventurer in need is useless at comparing tanks to healers
Again, the lottery was conducted on an external server. The external server did not communicate the data at the expected time, so it defaulted to no (you didn't win the house) even though you did.
It's not about booleans but about how normally-working code can be inefficient on a large scale (like an MMO) and create a bottleneck.
The issue was https://eu.finalfantasyxiv.com/lodes...d233a2ab8e17b6
Quote:
2. The lottery was conducted successfully, but the winning number was not communicated to the housing server
After investigating this issue, we discovered that, while the server responsible for conducting the lottery is functioning as intended, the results are not being properly relayed to the servers responsible for managing housing-related data.
Once this issue has been resolved, and lottery results can be relayed to the housing data servers without issue, it will be possible to confirm the lottery winners.
The lottery data that exists on the server responsible for processing results will be used to overwrite the housing server so that the correct data is reflected.
I am! And this can be how it works in some less agile companies. Given SE's rigid patch cadence and the size of an MMO project, it really wouldn't surprise me to know that that design was signed, sealed, and delivered months ago, and everyone involved was too deep in their next work to really go back and iterate - but it also suggests the outcry wasn't really big enough to force their hand. We've seen them move quickly when they think it matters. Unless there's precedent I can't think of for something like that already functioning in game, that'd be a significantly taller order than you might think it is. It's one thing to have an item step up from one amount of slots to another with no respect for its previous state, but just adding a third slot to an existing item without blowing up the item and creating a new one in its place would be a pretty significant systems-level change.. and with that comes a whole lot of QA.
I think what the jobs really need now is just inherent traits that give it unique buffs, movement options (like Dragoon Being able to jump higher than other jobs, might be fun for jump puzzles), and have gear that has specific traits only that job can use.
Or anything else at all, just something to give meaning to the jobs again.
Toxicity is Toxicity
Both can be Good and Bad.
I mean they can just put it in the shop, duplicate the item and call it "Augmented Ascendant X of X" and make it have three materia slots and use the same model, it's quite literally just adding entries to their database. Takes literal minutes to make such an item.
Then again they managed to screw up even basic database management with the Diadochos bug earlier this patch so lmao, can't expect much from them.
I like the forums lately because it seems like the majority of people have caught up to the way I've been feeling for a while. We don't really want to quit. We become harsh critics because we want S-E to do better and give us reasons to stay.
It's a flagship product carrying a brand that has delivered some of the greatest games of all time. They never should churn out the same gameplay loop for so many years as if it was perfect from the start. After 10 years of iterating the same patch cycle, they can do better.
Right, that would be the 'with no respect for its previous state' option, which is easier but more limited since it means you have to duplicate items every single time you want to do it.
Screwing up databases is so damn easy, though. The real question to me is how it slipped through QA's fingers - are they just skipping production client validation or is SE slipping last-second changes into the tables?
why do they put the live letters at 4am my time without subtitles
Now you see why I mentioned earlier that this post contained a similar talking point to Mr Happy's video.
Although this is a good point in some cases, it is not a good reason for the rewards to be unmodified. The arguments being made are to make it useful for gear, which is as simple as making it drop a right side or (maybe) left side augment once a week. These items already exist and require no extra development such as model or icon design or concept art.
If something involves concept art it makes sense why it won't change fast but this isn't the case when it's just adding augments.
That's the thing, this should be a simple check and shouldn't be done on another server that they query that's a spaghetti way of doing things...If the characters are objects you can relate the house object to the player object, doing it any way but than that seems very inefficient to me...I mean you can query SQL server data and have it read/ store it there but that's very inefficient for something that should be simple in practice...a lottery should be a boolean no matter how I look at it or at least baked into the server logic[so returning it is compatible rather then left unresolved]. You either win or you don't rather than take the time to query a server from...1..2..3..4..5..6..7..8..9...to how many total players, you see how inefficient that would be, of course they ran into the query stalling...(?) [That should already have been a concerning issue]. Assuming they are using IDs representing players to store information which again...they really shouldn't for something so simple. It's like if I asked you to keep picking specific socks out a plethora of cloths draws from the next door room with tons of different: socks, clothing, ties, suits, etc...[now imagine that expanded], but what should be happening is: the sock draw is already opened on the start and the person using the draw is you. (Understand the logic?) Of course big data takes forever to increment through. But, they shouldn't be querying data for something simple as a returned yes or no [true or false]. (There should be zero bottlenecking dbs, especially dbs on separate hardware that might have other information needed to be read in, for something that can resolved be relatively easily by relating objects). Normal programming would be fine because we already know the player exists and the house exists...You just make the connection, this doesn't query players but connects the action to the player or in this case the house to the player by functionality of the house thus being relatively instantaneous. <--- This is how many other games handle it. Assign function...pass character doing said function...Storing a bid might have some usage in a database but where it's just winners being randomized and bids fixed, bids do not need to be stored leaving the logic all binary in nature. I've seen games assign players an item that relates to that house indicating they have bid (This might be a better solution given if the server crashes or goes down for maintenance the bid remains...)Quote:
It's not about booleans but about how normally-working code can be inefficient on a large scale (like an MMO) and create a bottleneck.
I work with game data and server code and big data is used when it needs to be, a good usage of an SQL database on separate hardware for example might be for messages, maybe even item information(?), save states, etc Note---> [Things that are checked once rather than constantly checked]. If you query big data continuously for the heck of it, it might freeze table information. Although, I suspect I know what reason they did it this way...and it's not a good one, the Japanese developers/businesses like their spreadsheets so I would fathom a guess that they keep that formatting as close as possible by making or loading an existing id database revolving around those spreads(Which would account for bottle-necking a return value from big data) [Which is terrifying tbh]. Some good advice from my coworkers: If I'm waiting for your answer, your answer was not worth waiting for. <--- not exactly for coding or data querying but works hilariously enough for the answer to that but yeah....I'll finish my thought here. xD
MMORPGs often split things up into multiple servers so if one fails or equipment fails, the whole thing doesn't go down...MMORPGs are much larger than the little console programs, UI applications, single-player games or otherwise non-server applications most programmers are probably used to.
- Login servers.
- World servers.
- Area servers (ie. Central Shroud and East Shroud are separate servers).
- Chat servers (which is almost certainly how things like CWLS, CW parties, PF and FC chat function across worlds and why occasionally these have had maintenance).
All I know is that in a game of this scale, database access congestion could be a real concern and although they are never specific, they do keep hinting at either database or network congestion such as "the player only has so much data they can carry with them" which is why you don't carry the Chocobo Saddlebag into duties.Quote:
I mean you can query SQL server data and have it read/ store it there but that's very inefficient
I don't know why they did the lottery on a centralized server instead of on individual world servers, but I can see some risk mitigation in it. It prevents it interfering with the world's spaghetti code, it makes sure the lottery is synchronized across all worlds and data centers and frees it of delays that may be caused by the world's operations and likewise prevents it from delaying the world's operations. But obviously for all its risk mitigation, there are some risks it presents which came true.Quote:
Of course big data takes forever to increment through. But, they shouldn't be querying data for something simple as a returned yes or no [true or false].
I will admit I have seen all these things not be an issue on modern hardware (and this was years ago). So I don't really know. But from what they said, somehow they created a bottleneck even if it was due to an incompetent method of storage or working around spaghetti code.Quote:
(There should be zero bottlenecking dbs, especially dbs on separate hardware that might have other information needed to be read in, for something that can resolved relatively easily by relating objects).
I work with game data and server code and big data is used when it needs to be, a good usage of an SQL database on separate hardware for example might be for messages, maybe even item information(?), save states, etc
Yes I know, I'm speaking in perspective of someone who has worked on mmo code before XD.
When you know the mmo is experiencing massive amounts of people you want to keep it concise as possible...when you know something is binary in nature you make it binary...no super trick ponies needed. The issue comes with storing information... and I've seen items be the solution to that if only expected to hold a singular value that you want to keep saved. These items are usually dispensable to save it being written and kept. When doing large sets of information that you know is checked once the etiquette is to then think about having data tables.
Back to what's at hand, memory could be an issue but if they're having issues with memory to begin with...there's something else wrong. They either need to upgrade their server or figure out what logic or declarations can be simplified if it's garbage collection. I have seen lots of foreign developers run on a tangent on live code with ifs, comments, and blank space, it's usually the ifs that call their own functions that create problems but all of truly affects memory somewhat if it has declarations. If their conclusion cannot be assumed quickly i.e....do work code takes time to get to (this is less to due with memory and more performance) or they are experiencing high resource usage (this due to declarations or things kept), they should do something re-evaluation. Purge what is not needed.
I imagine the bottleneck they created happened because the promise was not met and when multiple asks on the query kept happening, that information wasn't readily available and returned nothing because the table was frozen. As the information is still utilized in the main code...If we assume large amounts of data...you're bound to experience issues especially when having to requery the same 20 million records when you already querying those records...and who knows how many fields if they're keeping it saved to the same table as available characters without it being partitioned (Which hopefully they didn't do this but you never know and since I have seen things horrible things from mmo developers in general it wouldn't surprise me if this was it.) [There's also what's being stored that matters too..] [Basically Database Overload] <--- Although that's just my guess. Could also be a combination issue where it's this mixed with poor querying in general. 20 million lines is a nothing query but..if it's being done multiple times at the same time could have issues.
The game has so many small issues that can be improved, and i don't see any of them adressed, or very very rarely any improvements.
It is probably badly coded in the first place.
Did the team got tired over time and all the past expansions, or is it the usual case of californian laziness where devs just do the minimum + focus on cash shop?
inb4 "If you don't like it, don't play" - This approach does not solve anything. Multiple people notice and point out it's on downward slope and the offered solution is to ignore it.
Ion winning by not making wow better but by sabotaging ff.
https://imgur.com/a/MyvG9SO.png
I agree that there is evidence that the devs listen to us and as a long time player, I'm grateful for the amount of regular communication we have, but there still comes a point where it feels as though we're being ignored.
Housing is my best example, and I am fortunate enough to have housing since I've played for so long.
There is no reason they haven't DIRECTLY addressed the inclusion of instanced housing, or what their long term goals are considering we are pretty united as a gaming community in being dissapointed in what we have.
There are other things as well but I wanted to give one example in case a dev sees this. I'm not asking for you to do everything at once, but I do find it tiring that live letter after live letter we aren't receiving anything of substance on the QoL improvements for the game.
We cannot go another 10 years on this foundation in good faith. We need the spaghetti code addressed, or for you to tell us finally what the limitations of it are and why.
I just want a clear roadmap of the journey you want to take this game on in terms of server stability, addressing all the system and QoL concerns that come up often and your vision for reaching that goal.
Maybe too much to ask but that would be what I want if I could choose.
Definitely makes me raise an eyebrow every time there's communication from the devs and the multi billion dollar company can't pay for a live translated restream or subtitles. NA has to rely on randoms from reddit, who are subject to mistranslation, for all of their information.
That's honestly the more damning thing is they can't be bothered at SE to properly communicate with their NA players so why would they ever pay attention to their feedback.