#server-plugins-read-only
1 messages · Page 22 of 1
No, your'e trying to explain to me how prediction and rollback work with local client, and I have worked on games where we disable that entirely when you're not a host, because we actually owned the entire stack and didn't need to do that.
MP = marketplace? That’s not the definition of supporting servers. Servers were within the release of the Java version
Multiplayer
Ah LOL
Well for me it’s a intended feature if it’s within the first release candidate
Which it was
Explain to me how you develop a game "client server" in which your clients were not delayed? - Fundamentally UNLESS deterministic (which is a different architecture) would give players a terrible expreience due to the travel time for the client to the server then back. This isn't something you can just "disable" it is a solution to a problem fundamentally with the architecture.
Should specify real time.
Delayed compared to whom? We are talking about single player. What are you delaying compared to?
"Client - Server" assuming your client IS NOT the server its self.
Again, it's a single engine. You don't need to receive data from the server. You literally render the server scene to the screen. If you're playing alone, that's all you need. If you're hosting, then you start looking at P2P vs server-client, but even then, in most games, the hosting player simply renders the server scene. There is no reason to run server stand-alone.
P2P? Huh? Last time i checked dedicated servers arent.... sorry i might be confused xD
oh right when ur friend joins ur world thats technically p2p.
Not really!
Soo what's UPnP then?
Yes, that is assuming the client is the server... (Not all servers are clients) - You are correct, the host is the server as awell as the client, but that is not the same here so in your statement NOBODY but the host would have a good time... We are talking about dedicated servers.
Engine or not doesn't matter.
UPnP is very different, but P2P is a specific kind of server architecture
Dedicated servers are when there is a server, and all the clients are connecting to it
we’re using UPnP (Universal Plug and Play) protocols to handle port forwarding automatically when possible - if that is unavailable, then we attempt NAT punchthrough..
that.. sounds like P2P to me..
We specifically weren't talking about dedicated server in the context in which you were replying.
You can use UPnP to open a port to a dedicated server
The fact that you're host doesn't mean that it's a P2P server architecture
Oh, then your correct! - If you make a single player game, we have no need to discuss Client Server architecture.
Host/Client still applies with the same rules of dedicated.
now that im thinking about it again true lol
P2P assumes both clients share equal (or mostly equal) responsibility in running the game. That usually involves both games running full sim and agreeing on how to synchronize that. Desync can cause either a reload from one of the peers or a complete disconnect. Most games don't do that anymore outside of limited co-op, because desyncs become progressively harder to manage with larger games. It's a lot easier for one client to become the host, that is take on duties of a full on server, and other clients to connect to it. Whether the host then "simulates" the entire server-client song and dance for their client is up to particular architecture. Some do either because it's easier to implement or for fairness. But because that creates a lot of overhead, and you usually have limited resources, for games where latency is not crucial, that's short-circuited and host simply has less lag than other players.
off topic real quick but i hope gta 6 isnt P2P.. it's a nightmare on GTA 5
It won't be
Sorry, should correct myself. I sure hope a game with more than 2+ BILLION dumped into it doesn't do so. 😄
anyway, cant wait to actually play hytale in 38 days
I'm curious, does anyone think that Hytale is making a mistake by only letting players play on the "current version" of the client, as they said?
I think letting players and server owners play under different versions was a strong point in minecraft, although that did present a few problems too. But the major advantage was that plugins we love did not break and become unusable when there was a new client version.
In minecraft there are many great plugins for 1.8 and 1.12 for example that only work in those versions and the authors have moved on to other things in life and so won't update them to the "current version". I'm curious what others think?
Early on, this might be a hinderance, yes. But once the API stabilizes, hopefully, backwards compatibility isn't going to be as big of an issue.
Why would you want players and server owners to be on any other version then the lastest it makes 0 sense and considerng how awful it was for minecraft to allow that, yes its the best thing that hytale can possibly do
It's interesting that you consider it "awful" and yet many people think it's one of minecraft's strongest points. But of course people will have different opinions based upon many factors.
i dont see how its one of minecraft strongest points lmao it allow servers to be on outdated versions and provide no real reason or innovation for servers. Its awful to do
Hipixel itself stayed on 1.8.9 for many many years precisely because the later client versions did not support all the work they did to build up a great experience for their players.
hypixel did it because of the playerbase wanted it and because of the pvp lol thats all
which isn't a good reason to stay, its outdated and there's 0 innovation
Exactly, they did it because of what the players wanted (but also to save themselves lots of re-development costs to change all their code for newer versions).
It's all about advantages and drawbacks.
What's the advantage? There's no playerbase split between each different versions.
What's the drawbacks? If there's any breaking change, then it requires mods to be updated to fix those problems.
Solution? Make as few breaking changes as possible.
I feel like Minecraft is not exactly the best example because its modding is not as good as Hytale. In Hytale, mods can be "built-in" to the server whereas Minecraft they were not. If the Hypixel 1.9 PvP update situation had happenned in Hytale, they could have just made a mod to recreate the old PvP that fit their game. See the difference here?
I think it will be fine to be honest. Hytale's modding is very extensive
If it really works so seamlessly them I agree, we will have to see if they can pull it off.
I don't see why it would not for like 99% of mods to be honest
Minecraft isn't really built for modding at all. "Recent" versions added datapack and resource pack support, but a) that's almost an afterthought for how Minecraft was built, and b) That still doesn't have nearly the extent of flexibility of plugins, which had to be retrofitted into the game by the community. When your main tools of getting mods into the games (Forge, Fabric, etc.) are maintained by different people than the game itself, and also with Javs vs Bedrock disparity to add on top of it, is it a surprise that version updates break mods so frequently?
Exactly. And anyway, at some point, if breaking changes are required to modernize, then that's fine too. They could also just warn modders about any breaking change and have a plan for modders to be able to fix their mod even before the version is released.
Java is almost an abomination however if Notch never coded in it we would never have gotten such an easy modding system we have today. It was worth being made in Java for all the custom content we got over the game's lifetime
But yeah. With modding API being introduced, and once mods have no reason to rely on code injection and the like, version-to-version stability of mods should get a lot better.
GTA engine is C++, and modding flexibility is almost better than Minecraft. The language doesn't really matter. Yes, the fact that Java is easy to decompile makes keeping up wiht version changes a lot easier, but with a stable engine and active community, we'd have much the same effect with any other language.
I will pray along with the rest of you that server plugins will be timeless and work with all new versions, and won't need the creator to regularly update them, but I have a sneaky suspicion that it will not go that way. I actually hope I am wrong.
They already said there will probably be a lot of breaking changes at the beginning. And that's normal, the game is in development.
Yea, that's what got me worried in the first place.
It'll be chaos either way early on.
Once the API is more stable, they can very well make updates where they keep old stuff and mark them as being obsolete features for some tiem so that modders can modernize without the mods breaking. It's very common in software development with language and framework updates
Also, just because there's no official support for running older versions, given that it's still a Java core, I'm sure there will be a backlog of previous versions somewhere. If it becomes critical to freeze a given version, people will find a way.
Uh? Can you develop pls
What do you mean?
What do you mean by "freeze a given version" ?
I mean run an older version as a server without letting it update.
I don't think the client would like that. Then again, yeah maybe
You'd almost certainly need to have both, with each one modified to prevent updates, yes.
Yeah. Unless there's something blocking that, it should be doable (and even then)
I don't really see why devs would block it intentionally. I can see why there would be an automatic verison check, but I don't see why there would be additional measures to guard against version check being bypassed. This doesn't seem like the kind of a project where it'd matter. Of course, some reason to disable it can come up in the future, and then it could turn into an arms race, but I hope that's not the case.
As a general rule, whatever devs put into a popular enough game, players will find a way to break it. So long as nobody is using that to cause harm to the project or other players, the best response is usually to just let that happen.
This is why I considered minecraft's decision to allow the client to run past versions was a smart move (even if it had some problems), because then they could avoid/bypass these sorts of issues. There have been many mod/plugin creators who put thousands of hours into beautiful creations, but then they got married, got a new job, or passed away etc, but people still wanted to play with their mods/plugins. And so if Hytale comes up with something new in the latest client that breaks these things that thousands of players have come to love, you will find that someone will find a way to make an alternative client, which is not what I'd like to see be required to maintain compatibility.
That's kind of a problem with all games, though. Not just mods. There are people who work on game preservation, but it's a lot of work, and it's not always possible.
Not to mention games that never get released...
I don't mean to make it sound too much like, "That is just how things are," but it's a greater issue than just mods. To some degree, games will always be a transient medium.
It will be fine 🙂
Good plugins and mods will keep living on
Maybe in this age of AI coding, people can take mods/plugins from dead developers and have them upgraded to the latest game version with less work.
Lol, AI won't help with that
But if a good mod gets broken and stops working and the modder stopped working on it, then another one will usually pop up
As much as I'm not a fan of Java, mods being written in it means that decompiling a plugin is just as easy as decompiling the game. How readable the decompile code will be kind of depends on how good the coder was at keeping the code modular and clean, but if all you're trying to do is update it to a new version, it should be doable more often than not. Minecraft had a lot working against it in the early days. Like the way block IDs work, and chunk ordering getting changed part way through, meaning breaking changes were usually REALLY breaking. I don't anticipate things will be like that with Hytale. I think mods will be more easily fixable.
Also have to remember that most mods will be made with the modding tools:
Assets editor, block bench and node editors.
Most mods will always be fine because the tools will always be too. It's really just the more advanced java stuff which I would expect to see better support from people doing those
I'm glad the rest of you are so optimistic. I guess I just have a knack for seeing future stumbling blocks. I really do hope that Hytale has thought of all these things and came up with a bulletproof design.
Again, there's no bulletproof design. Sometimes, you just gotta take the bullet 🙂
Didn't they say that they want to have only 1 version only?
As in, new versions of the game supporting older versions of the plugins.
I've been at meetings where that was the decision, let our users take the bullet. Sometimes there is no kickback, and other times, well ...
No offence but shouldn't the devs keep up with the API updates? just like discord
Plugin devs? Theoretically, but they're not working for anyone. What if they don't want to? Or are busy? Or just moved on to other things?
Yes, but some mods might be real big and require a lot of changes. Having a lenient API is good 🙂
Honestly I don't know. I always update my stuff
That's fine. But some people don't.
Wishful thinking we won't have breaking changes
I assume it will be like any other famous API and just go with experimental and deprecated features for a while and then purge the people who refuse to move on
And many devs start out committing themselves to always updating their mods/plugins to the latest version, but life goes on and sometimes you decide that it's time to stop updating because it's become too much work, or you retire etc.
That's the point in time where they should make it open source
Anyway, I'm going to bed. Have a good night guys! Was a pleasure discussing with you
Yeah exactly, we should count the fact that sh*t happens for 3 to 6 months. But if you can't commit just let it go, why should a company WORK FOR YOU, when you just don't want to do the work and update?
I don't know what company you mean. Most nod/plugin devs are making their creations just for the love of it, and some maybe make a little money from the stuff they create. If they feel it's too much work to maintain their own creations, no matter how many people love it, they have all the right in the world to just not put in the work to update their mods/plugins to the latest game version, even if they are abandoning their followers. And sometimes they just die. We all know of a few of those.
I think it's quite simple really. If you don't update your applications then they will stop working in the future.
which is why you should make stuff opensource
if you cant or dont want to maintain it, at least let people who want and can continue it
You don't have to think in terms of plugins. Generally if you start a project and stop later, it will die
Yeah, but I can run a game from 70s on my PC, and it will work. That's kind of the point here. Work can stop on a project, and it can still be perpetually functional if all pieces of it are no longer changing.
I agree
I don't think backwards compatibility is something they should strive for before 1.0
But if there's a breaking change, you gotta update
I really wish that was what people did, but I have many old apps that I love that have not been open-sourced. And even if they are that is no guarantee that someone will take up the project.
What do you mean gotta? Like, who's forcing it? The problem is that there's a plugin that people want to keep using it but the dev is unable or unwilling to continue work. That's the situation.
I still have a few old apps that I have to fire up a WinXP VM to run. But at least I can do it if I need to.
It happened in bannerlord. Everytime they pushed an update mods broke. Who's fault here?
The the bannerlord developers don't advertise API stability, there is no fault for breaking changes.
Most of the time, development velocity and flexibility is more important than API stability. Especially during early access
We should be fine after EA then
Yeah there's a long time to go and the way they commit to being mod friendly, I'm certain they'll have good solutions for this in the future.
haven't they said client and server versions are seperate?
there's only one client version ever of hytale, and it can be compatible with any number of server versions as long as their changes haven't affected the client. so you can just keep your server on an older server version if there are breaking changes to the api that don't break the client. you're only forced to update if the client-server interactions got changed in some way. and regardless of all that, they've said your plugin will stay compatible without any changes if the newer server version doesn't touch the same code you were working in (so no need to publish a new version with new metadata that says it's compatible... ehm minecraft)
Will Hytale not support playing older versions? I remember them saying something about the client only being able to play the newest version
yes only the newest version of the client, but unless the server protocol changes (which it might in early access especially) it should work with any server version
Might become a problem, especially after early access
I don't see how, the only client/server versions that would be incompatible would be ones related to security or network improvement
Bcz then mods will need to constantly be updated (or just die bcz u can no longer use it), vs Minecraft where users very often play the previous/older versions until the mods port
mods aren't on the client anyways, so they don't need to update. you can play on any server version you want
I assume the protocol will be breaking every so often, especially for adding new kinds of abilities
no, that would be sending new things to the client to render from the server, just like mods will be doing, doesn't need a client update nor does it make incompatible
protocol as in changing how things are sent to the client, changing the api that sandboxes the client in a breaking way, etc.
No, I mean new kinds of abilities for the protocol itself
Like maybe more options for the transfer packet (or whatever, the Minecraft protocol changes almost every version)
And it shouldn't be that hard to have a basic version selector (can be as simple as installing the entire game into a different directory, like Hytale/v0.5 Hytale/v0.6 etc) then the launcher has a drop down to select the version to launch)
you can very easily design those in a way where if a newer client joins an older server, it just won't receive those packets and won't do whatever the packets were going to do
not every single change to the protocol has to be a breaking one, but if they for example decided to scrap like half of their networking and rework it then sure it would break. that code doesn't change every so often
idk I think you should trust that there are lots of modders in the hytale team, and if there's anything modders hate; it's porting constantly instead of actually working on the content
In my opinion, giving users the option to play on an outdated version for an extended amount of time would be a mistake. Servers would petrify on an old version and force their users to downgrade and then demand support from the Hytale team for their outdated setups
That's assuming it only adds capabilities, but it can remove/change them, like imagine if the hunger system was completely reworked, so instead of a hunger bar, now u have 5 bars each for a different kind of food (or whatever, just an example)
Well now all the plugins I bought for my minecraft server are going to waste
and that would be possible without breaking. the actual hunger calculations are done serverside the and ui is sent to the client from the server side too. basically, if it's possible using a mod then it's not a breaking change
Not really? it's obvious that older versions won't get support, this is how Minecraft works, u don't see updates to versions like 1.7.10 or 1.12.2 even though some really good but old modpacks still use these versions (due to certain types of mods)
imagine if one of the hytale client versions had a huge security problem, you really want to make it possible for the players to go back to that version? it's an intentional decision to keep it a singular installation and a singular client to maintain quality
you're confusing minecraft versions with client versions, infact there is literally a minecraft plugin that allows the newer clients to connect to older servers. so it is more than possible on hytale to play on an older server with a newer client.
Sure, it depends on how hardcoded stuff is, and what can be a mod vs can't, my point is breaking changes to the protocol will happen, which is fine, but I see no reason to not just allow players to keep older versions even if unsupported
well there are reasons, which they see
but again if it's possible with a minecraft serverside plugin to make newer clients connect to older servers, it'll be more than possible on hytale where most of the code is server authoritative
if you have a theoretical version of a 1.12 modpack that has newer client code versions of the mods, you can connect to said modpack in 1.21.1
ViaVersion has a bunch of issues, for example shifting in newer versions reduces ur hitbox, same with swimming changing the hitbox, if u try to do that it won't go well
Won't there be security issues if they don't update client
yes that's because minecraft's client code has assumptions and power, it's what makes hacked clients possible because the code is not server authorititave. it is on hytale though :) hytale servers are the ones that tell you everything, the only thing the hytale client does is just listen to everything the server tells it and renders it and sends signals back according to your input
To be fair to ViaVersion, it has to do a lot of heavy lifting to make things work, and does a pretty good job. Heck, even clients that don't yet support negative y altitudes can connect to a server that runs 1.18+.
In Minecraft, server and client always update at the same time (barring some small hotfixes here and there, but the version will still be bumped). I am not so sure if in Hytale this would also be the case. The client could run on its own release track from the server.
What if there's a sandbox escape in version 1.0 and people stuck with that because of servers and hytale doesn't update it?
also in this case, we're talking about connecting a newer client to an older server. this example is connecting an older client to a newer server, which is not possible in hytale
Sure, but I think they should at least allow players to keep previous versions, especially breaking versions (security/small patches r fine to keep as a single version)
And clearly label it as Unsupported/Insecure but allow preservation
and why do you think that after literally everything I just told you
you had valid worries, and I told you why they don't apply to hytale
No, I mean newer client to older version, bcz the server thinks u r phasing through blocks (and rolls u back) causing the client to be very glitchy
the client will have its own release track yep, and it'll always be a case of connecting newer client to older server not vice versa which is way more reliable in viaversion anyways
ah I see, you're talking about client prediction. yeah hytale mods can't even do that since they're not on the client and it is very safe to assume the hytale client won't either since it's not meant to contain the game. there is serverside lag compensation to account for that though
client prediction is when the client expects something to happen (like the hitbox decreases) so it renders/does it early so that you as the player don't notice the lag of having to wait for the server to confirm the action. since hytale mods are all serverside, they can't make the client predict anything and thus is it more dependent on internet connection, which is where the serverside lag compensation comes in
it would be monumentally stupid (and they know that) if the hytale client only predicted vanilla things and not modded things, so hytale clients will definitely not be predicting anything to keep them separate from servers
I highly doubt the client won't do any computation regarding movement, bcz it'd look extremely jaggery to try and get positions from the server in realtime
Most multiplayer games compute movement on the client, and then the server rolls the client back if the real position and the client position are too far apart
search up serverside lag compensation, it's what helps against this. maybe I am wrong and there is some prediction, and that would be where clients and servers start to have problems. however there are other ways to fix this like by asking the server when you first join if it should predict x and y or not
I'm not aware of any team member who's talked about client prediction, but they've said that clients and servers are on separate versions so they've definitely accounted for it in some way
Assuming no prediction, that means that for 30fps, u need to get positions from the server every 30ms, or 16ms for 60fps, and a round trip usually takes 100-300ms (depending on how far u r from the server)
I think the pros of a singular client outweigh the cons, and if anything in the absolute worst case scenario you can add code that only runs when in an older server. it's like if minecraft made sure that the server was above 1.13 before making the hitbox smaller
did you search it up or do I have to explain it here?
I am not against that, I just think there should be a way to preserve older versions (even if unsupported), even for the sake of trying out a really old version from 5 years ago (and servers are just jars so it's easy to just keep multiple versions)
well the last thing is that the servers are indeed just jars, if a very old version of hytale used an older protocol no longer compatible with the new game. you can just rip out all the content from a newer server jar and put in the old content and adapt it to work with the newer protocol and play on it
look, there's a solution for literally any problem you encounter and having one client is too much of an advantage to pass up atleast that's what the team thinks
(also you'll probably be able to play on older clients since they're just code, you just won't be able to log in to your account on them)
the word I was gonna use is banned in this server lmao but I think we all know it's possible to play minecraft without an account and join offline servers
That doesn't fix the problem it only corrects it on the server-side to be more fair for the player
For movement to look smooth, u use interpolation, so if u want to move from 0 to 10 in 1s, at 3fps, u will be at 3 points, 0, 5, 10, but this doesn't look good, so u interpolate 33ms for 30fps, or 16ms for 60fps
U can't receive updates from the server this fast, so what basically every game does is calculates where u will be on the client side, even if it's slightly off from your real position on the server
Then the problem is, that your actions that u see on the screen, are slightly off from the actions the server calculated, so the technique u mentioned for lag compensation fixes it
yeah I'm unaware what they've done for prediction I only know that mods can't send executable code for the client so even if clients predict movement, it would look stuttery if a mod is changing your movement. it's a bigger issue as a whole, interested to see how/if they've tackled that
But u still need to do client side predictions, which can be as simple as interpolating an entity based on it's currently known velocity (so it looks smooth)
The problem with that is that if the velocity has changed, and u didn't get the newer one yet, the position it's at is off from the real position, causing desync
The question is how u correct it, Minecraft just rolls u back to the actual position if the 2 get too far apart, or now u also introduced me the lag compensation solution
It should only stutter if the mod rapidly changes ur velocity (if it's gradual, the ~200ms delay won't be a huge deal, if it's significiant, then the margin of error would be noticeably large)
Or, crazy suggestion, we don't have mods that alter your movement in very unpredictable ways. I might lack imagination, but I do not see what kind of mod would require full control over the player movement in such a way that the server would reject the movement outright.
I wouldn't like it even if it wasn't a huge deal, they can definitely think of some way to make clientside prediction possible on the serverside. it sounds counterintuitive but I can think of a serverside api that doesn't give you power to do anything you want it just allows you to do calculations and the code that leads to those calculations is sent to the client when they first join
The server would probably be somewhat flexible with movement. Even in Minecraft, you can permit via plugins illegal movement (mostly as you can control what packets are sent back from the server to the client)
that's one of many examples, but no I can and I have. for example, I've made a mod in minecraft that anchors the player to a certain point so if they're moving away from that point they get progressively slower
Yea, the only way to prevent large movements being stuttery would be to somehow tell the client how to compute it themselves
Sounds interesting, but how would that be "illegal movement"? We can surely apply speed modifiers to a player depending on their location
yeah that's what I'm saying, you have validation before it is sent to the client to make sure that this code does literally nothing except calculation. can't tell you I know what I'm talking about though, haven't tried doing that before. doesn't sound impossible though
those speed modifiers that get applied would be subject to latency
Things like custom jump curves which can be used for things like vehicles/mounts etc (imagine a mount with a specific jump) the velocity will rapidly change, especially at the point when it changes from going up to going back down due to gravity
And u can imagine mods like flipping gravity or side gravity etc
Assuming a mod does allow some control over the movement of a player, that won't be an issue right? So long as the server agrees with said movement based on e.g. you being on the correct mount.
literally anything a mod does at runtime and not just when first starting/joining the server will be subject to latency, I'm not sure how much of an issue it will be and I doubt there's a solution in the current early access but I hope one will be reached eventually
U can't just send code over the network like that (I mean u can send source code or bytecode like lua/js/java bytecode/C# bytecode, but it'd be extremely unsafe)
yeah exactly I would assume it's unsafe too, I'm not too knowledgeable on this so it was just spitballing
I don't think Hytale will have a system like that, especially not anytime soon/early access
but it doesn't have to be straight up code, you could tell the game which numbers you want the client to use to calculate, then which operations to do on the numbers, would ensure you're not doing anything crazy (maybe limiting though)
The less dangerous way of doing that would be to have a command protocol, where the server sends commands, and the client executes those commands, then commands can be like get this property of this entity, normalize the vector, scale it by (get this other property from this entity) then set the result into the "velocity" property of that entity
Then the client basically has a way to compute velocities
would just be sending a stack of numbers and operations that get substituted at the client
But that'd be fairly sophisticated to implement
yep! exactly
Yeye, but now u r basically inventing a sort of stack based programming language, def not something you'd expect them to have in early access (or at all)
oh yeah no they've already said the sandboxing for the client isn't there at all rn like even shaders aren't possible
And the less limiting version is to run a full blown programming language like lua or javascript but only allow those to call specific external functions (to get properties etc)
And that's basically how browsers work, they run javascript, but js can only really interact with browser APIs like modifying the DOM, it can't for example open a file on your computer, which is why things like local storage exist, a special API to safely allow websites to store persistent data
Ye serverside shaders will be pretty unsafe, u don't want to give servers full control over ur GPU, they'd use it to farm crypto or something
Will be safe the way they do it
That actually is a crazy suggestion lol
A hybrid approach where you can toggle clientside movement would be infinitely more applicable
Minecraft is something that does this already, and I don't see why opening that up to developers via a command sent to the client would be deeply problematic
If you want to make a flight simulator in Hytale, it shouldn't be infeasible to outright control the player movement server-side and just take the user inputs
When starting the server, again VDS (are we setting it up from a virtual server?)
What do you mean?
Forget it, if you had a brain, you'd understand.
Well then!
huuh
It just sounded like a very simple question, and such a question doesn't align with how competent you were speaking previously 😂
So I figured I was misunderstanding your question
if YOU had a brain you'd write it in proper english
I'm Turkish, my friend, the translation says this, what should I do?
that's not an excuse for being rude lol
Stop comparing hytale to minecraft
Minecraft is SHIIIIIIIIIIIIIIIIIII, there's no comparison
Everything you use to create is third party. Nothing is official
It's a game with user content and networking, among other similarities, it's not completely incomparable
And a lot of the references to it here are just to provide examples to explain better
Everygame is user content and networking. But no multi billion dollar game tells you to reverse engineer the packets to create your own thing
then don’t insult people for not understanding a sentence you didn’t even write clearly
sorry
all g
I am a little biased hater @sharp lake I am sorry. I will stop baiting
You can run server from a VPS or a VDS if that's what you mean
But theoritically you could access old version servers unless there's been breaking changes
That would def be nice...
I hate Mojang with a passion at this point lol
It's just that most people here have Minecraft experience, and I'd imagine a lot of the Hytale devs do too
It makes a perfect example for explaining concepts
The reason hytale was made was because they hated mojang's rules and programming. They know the struggles that's why I don't expect anything simular with minecraft
They could artificially block it, but you can just unblock it too so
They know what it's like waking up and have to rewrite your whole codebase
That is a shallow take, considering the styles and genre (voxel game) are obviously similar
If there are no breaking changes, I don't see why you wouldn't be able to connect to a slightly old server
That's literally the reason dawg 😭
They were limited by mojang and wanted to create their own thing
I am not saying that Minecraft/Mojang is the reason for Hytale, I am saying that there are obviously some simularities between the two games
plot twist they're obfuscating their classes n shi as we're speaking
Didn't they say it would be unobfuscated already lol
As for needing to rewrite their entire Hypixel source code due to the EULA changes, also unlikely
wot isn't it?
well damn they did NOT understand our struggles then
Ah yes, I mean it isn't. Or well, Hytale was created as a response to the EULA changes, I am fairly certain they acknowledged this way back when
the first 3-4 months will be dedicated to mapping their code
They will open source the server eventually but for the start, the jar we get will be unobfuscated
I think you have this backwards
Obfuscated means scrambled, basically
What's going on here?
Also, the server code will be shared-source, not open-source.
sso minecraft is the reason for hytale after all
Indeed
I am not saying it was created as a FIGHT against the EULA. I am just saying that they were so limited with what they could do. Programmatically and money wise they decided it would be better to make their own game instead
^
oh okay i was just getting confused lmfao
So if it's unobfuscated, it's unscrambled
That means no mapping needed
Also over the years, minecraft deciding to break the law and ban java players from profanity, guns and gambling while allowing it on bedrock. It sparked the fire for everyone else
wait my eyes are playing jokes on me, i read "unobfuscated" instead of "obfuscated" (Didn't they say it would be unobfuscated already lol)
What are we talking about here
Pff some wild claims being thrown around.
McJeff you are not built for this
I guess I am not
Modding blog post:
We commit to releasing the server source code as soon as we are legally able to. Expect this within 1-2 months after release.
In the meantime the server is not obfuscated, so you can decompile it easily.
Very interesting
To be fair, Mojang's EULA has been entirely unenforceable for years, legally speaking
It doesn't stop Mojang from banning accounts and server domains though
Can someone educate me on what you can do with the "Server source code" ?
It means everything
nah they just didnt like that new EULA forbade them from making money through bad monetization
You don't have to guess blindly at all, you can just read and extend the code easily
Yeah I am talking about what you can build with it
It just makes everything easier
anything really
Like I have no idea what you can build with an open sourced server code. How is it different than the official hytale api for game devs or sum
- Look at how the server works, how the api is implemented, and what it does in the background (especially useful since the early docs will be very sparse)
- Extend it with custom functionality beyond the modding api by decompiling and recompiling it
Ah I see... well I can't wait to see what you guys build with it. I will simply use the server api for minigames
The API is how you interact with it. The source code is what it interacts with, and allows you to see what it does.
And well, you can extend it, so long as the server<->client protocol allows for
The source code is basically defining the API
So you'll know exactly what's going on when you use the API
And you could even extend the functionality of the API with it
Having the code open like that also means the community will be able to do "merge requests" (ask to add their own code) for optimizations and features
Ah nvm. I missed the part where people won't be able to use it alone, perhaps contribute?
Yeah I see now lmao I am lil slow
If you mean outright duplicate and distribute it with changes, almost definitely not
They would be using different language with that intent
If you mean private "forks", I don't see that being totally banned
If you wanted to compile the server locally with an extended API, and just run it
I doubt they'd send you a cease and desist
I don't think they would indeed, otherwise they'd see no need to make it open source to begin with if you aren't allowed to tinker with it
You want to use plugins when possible obviously, but if looks like there will be options
I don’t remember who but I swear someone said you’re allowed to just do whatever you want as long as you don’t touch the protocol, so as long as it requires a Hytale account and is compatible with the Hytale client then you can just rip out the entire code if you want
I also imagine plugins will be able to mixin and all that
You could be right, that would be extremely generous of them though
I also agree that plugins will probably be able to Mixin, especially with the state they plan to release in
Mixins will allow us to work with code that hasn't been given an API yet, and they admitted it's not fleshed out
I trust the hytale dev community will carry the server protocol. Bunch of high iq individuals
It would be pretty annoying if we had to patch the source directly instead of using mixins
But either works I suppose
You can make your own version of server software, it just has to respect Hytale's protocol (QUIC). So basically, the server has to work with the client. That's all
We'd just end up distributing a mixin patch for the server, and using them anyway probably 💀💀
Either would not work cause I’d have to download patches from people and take out their files to put in the server lol and then recompile
awesome
You'd only need the one patch for jailbreaking
I guess so yeah, mixins will definitely be allowed though they are modders they know invasive
is there any information how much will we be able to play with movement, camera and custom UI elements (or hiding some default ones) on day one/shortly after launch?
Hey, respectfully, don't ping me for a stupid ragebait about a message that isn't even recent
ok sorry boss
No super concrete information, but one video showed the camera in a fixed top-down-ish perspective. So stuff is possible, just unclear how easily accessible.
I'd assume the bare minimum which is controlling the velocity, and also pitch/yaw for the direction u look at
If there is anything above that, probably not bcz it's early access
yeah, the old trailer video, but i am not sure how much of that could we actually expect...
Well the tools in creative mode have fancy camera movement features so that could definitely be used as a starting point.
that would be great, bcs the fixed top-down-ish perspective is something i would be looking for since its missing in minecraft and could be used for many cool mini-games.
This community is so toxic and the game hasn't even released yet
Idk the only person who has stood out to me so far has been that guy LOL
Everyone else has been mostly nice
Rip Kyren though who got another ping over this lol
Referring to me?
No, yaga
Doesn't appear to be a dev, yet is super active in here and very 'matter of fact'
Oh wait I thought u were the guy who pinged me lmao
noooo
I also had another guy tell a bunch of people to "don't trust @prisma agate " (and pinged me every time for that) just bcz I disagreed with his opinion lol
Hopefully it's just outliers
Oh wow another NixOS user, that's actually insane LOL
I actually don't like it and I think I'd be switching eventually (when I am not too lazy)
Oh darn, fair enough
I haven't had to touch my system in months basically
I kind of crunched it out with my initial attempt
The only thing I wanted was to define my packages in a text file (so I can easily uninstall and not forget random stuff I downloaded)
I wanted to programmatically configure and install things across a few different machines, so it's been working pretty well for me
But the whole issue with not using FHS means a bunch of software can't find where libs r located, whichcauses me a heck ton of headaches
Anyway, you've been here a bit longer than I have so hopefully I'm not just sheltered 🙈
I haven't really found that to be a problem because you can ELF patch, or run it in a FHS like via steam-run
And then of course distrobox, as a last resort
Idk I just always get the weirdest issues that take me ages to solve, like I had no emojis for a few months in firefox, and then while trying to fix something else I found a config option to fix that issue, which is just so annoying to deal with that
Not sure to which distro to switch tho, I'd probably go with arch bcz arch wiki is really good so if I have issues it'd be fixed easily, and AUR is very mainstream so most packages will be available there and updated, the only thing I don't like is it's not stateless, so I'd have to touch stuff outside /home and I'd also need to remember what I touched (so I can rollback after messing around trying to fix something)
So I thought maybe openSUSE tumbleweed, but zypper doesn't have many packages and also kinda sucks
Chat is debian good
Have been using it on my servers since I started managing servers 15 years ago
Debian is fantastic! - Very robust, very stable!
Run typically run Ubuntu (Debian based)
But for pc?
Ubuntu maybe best if your looking for games, or anything relavent to gaming hardware.
You do have PopOs, Ubuntu (Some Thirdparty packages that many others do not have)
But isn't debian enough
There I'm using Windows + Ubuntu (via WSL2)
I prefer this setup much more. 😄
eh i just go straight debian
ubuntu feels kinda bloated these past few years
everyone trying to switch to wayland is also making it kinda not the best experience
Well everyone is trying to make Linux the gaming OS, which needs a ton of TLC before I jump ship. 😄 - I rather not tinker with my OS, its my tool not my job. 😂
hey tinkering aint a job its a hobby
You're right! - I just have a different hobby. 😄
fair enough
I however would love a linux based or similar but with the gaming support windows has. I develop games, and the tooling for windows is just incredible. 😛
well i can tell you for sure current steamos aint it, i tried
No matter how many YouTube videos tell me I am wrong. Your statement is very much correct.
like it runs games well, and has KDE on arch
but the dekstop experience there is miserable on multiple levels
windows still leads the way
OH my god! Yeah Linux has a terrible user experience!
I only use it for servers
Editing my messages are my way around the cool downs. 😛
I don't think many game developers care about pushing to linux
It's way better than windows it just can't be a gaming os
Not the big ones, but I do. I is only difficult due to Linux being a hodge podge of differences.
steamos is great gaming os, but just not great desktop os
all my games run, and in some cases better than on windows
I haven't tried it but still I don't think it will be enough. Also some games just don't run because of kernel ac
This is very common! - However, I play exclusivly multiplayer. 😛 and many games that I play either do not run, or can't rely on Linux due to the freedom of the OS.
i dont run games with kernel anticheat
proper cheats can still go around kernel ac, and proper anticheat should live on the serverside.
There is no "proper" cheats, it an ever changing minefield of attack and defense. Cheating will never be a solved problem. 🙁
which is exactly why companies who try to invade even further into my pc wont get my money
Yeah, I defintily understand this. 😄 - I agree with ya.
anyway nice chatting with ya all but i should catch some Zs before it starts getting bright outside again
cya
See u
I am able to run almost all games I want to play and I dont use Windows. I mean pretty much all of the games would be working on Linux if the developers didnt make a choice to break that by adding some invasive kernel anticheats.
I just recommend Fedora and Flatpaks usually 🙈
I haven't touched a Windows PC within the last 8 years lol
There is only one game in my library that bans Linux users and it's GTAV, which allowed Linux until recently
I've never accidentally bought a game that was unsupported, and I've never checked for support before buying
I wish there was a way to completely remove windows but I need it for some stuff
Kernel level anticheats just keep games cheaper currently. Having a fully staff team to counter cheating is expensive. - However I agree with your sentiment, as much as a I hate anti cheat I feel it is still current a nessasary evil.
Exactly, flatpak is just like the Microsoft Store, but one that actually works well and its great user experience 😄
The last time I tried the Microsoft store, it was like a billion years ago 😂
I searched "Firefox" and I managed to download a book about Firefox instead of an app 😭
Firefox was obviously not on the store at the time
Funny actually I only swapped to Windows in the past 8 years. before then 17 years in linux! 😄
ehm noo? I think its just game companies not doing a great job on the serverside anticheats.
That's what he's saying, it takes effort/money to do a great job
Not everything can be serverside anticheat. You do require a level of client protection.
I think that one is more debatable
Well, depends on what you consider anticheat - Obfuscation is a form of client protection. Handles protections against memory sniffing for leader board and scores as an example.
Kernel AC is the easy & cheap workaround for them
Anti-cheat is just anything that prevents people from cheating
Modern cheats don't even need memory access anymore
You have to detect behaviours, and there's no point in even trying that on the client, due to how it can be circumvented
When you look at Minecraft, all that the servers can do are serverside anticheats and basically the only really undetectable cheats are ESP, freecam and other changes to rendering, movement and combat cheats you can in most cases detect.
Waiting for CS2 AC for years
And is considered one of the most cheated games. (No idea how to reword this. :D)
Imo it's way more easy to detect and fix them because they are written in java and it's easier to RE them. Also they just abuse packets mojang didn't bother to check
Language is absolutely not a factor here
I recently inspected an AC for minecraft and it was just doing this:
If flying -> flag
It's a difficult puzzle on its own
Have you read the actual built-in anti-cheat? Lol
?
Minecraft has an anti-cheat
I'm talking about plugins man idk what ure talking about
you mean the disconects for flying on vanilla?
I don't think Java has one, but Bedrock does! If Java has one I haven't heard of it at least.
Flying, movement, block range, etc
Also language is a factor JVM is much easier to RE than assembly
Java has had one for a billion years
The "moved too quickly!" is anti-cheat
Among other things
Yeah, that is most definitly part of an anti cheat.
It's an extremely generous anti-cheat, but it does exist
There are some collision check related things too
Are you guys going to work on an anti cheat when you make a server?
We don't really know what the scope of hacked clients are going to be like for Hytale yet
Yeah, but these are the same systems CS2 employs, but we are talking about the wider world of cheating - Many things can be circumvented which is where serverside isn't enough. - Again, harder to do but just as easy to download an app somebody else made!
I hope it's hard making external client mods
Well mods here for the client would be called hacks. 😄
It's only a matter of time
Client side anti cheat is inherently flawed. Just infiltrate the guards and you're good to go. Cheats have gone from manipulating RAM and hooking functions, over elevated processes and drivers, over DMAs to screen capture and hardware mods. Some inject back into the game via popular overlays like discord or steam because those never get scanned. The only long term viable thing is server authoritative + behavior analysis - i have to agree with MMOs and Valve on that
im running fedora kde linux and honestly run anygame without vangard or kernel level anti cheat fine
Nothing will be enough unfortunately
And client-side anticheats don't change that, any advanced client-side anticheat is going to be a waste of effort
Try openSUSE tumbleweed
you can user linux for gaming just depends on what gamer you are. as a rule even on windows i never cared for kernel level anti cheats. every game i play runs on linux. and steam now leading the charge with proton and pushing steam os means this only improves
It's a very stable rolling release and it's for normies (not much tinkering)
I will take a look!
linus torvalds uses fedora linux so thats an edorsement if ive ever seen one
Alternatively fedora is also good for no tinkering, but it's not a rolling release (so less newer packages)
Was just about to say that
To clarify why it's a waste of effort, it's due to the circumvention that is inherent to trying to control what people do on their computers
If they don't control the machine, they can't easily circumvent it; aka server-side anti-cheat
Ryan is probably not reading it but it is on his list to eventually add "sub levels" which would allow you to make Movable structures like ships, planes, etc. This is a future work item and currently has no timeline. We really love that concept and the possibilities from this.
😮
the amount of automation this could create as well
Create mod right? Also trains 😮
What's the context on this quote?
Can we get relative gravity/forces? 😄
Hi team! regarding a "Create Aeronautics" style mod in Hytale: Can we render local block grids with custom transform matrices (moving structures)? Also, does the engine support client-side prediction/interpolation for custom server-side physics?
Relative gravity would be so cool
oh my gosh don't ping them, i just wanted to know where it came from
I didn't ping them on purpose I copied the tweet 😭
Twitter, got it 😭
I can't post links
doesnt help discord auto pings when replying
slikey replies
every day i come in here hoping something releases as to game code , or high level architecture. tomorrow for sure.
LOL don't hold your breath
It is going to be a field day on release
oh shi
Haha no worries. Just changed my Twitter handle to @random magnet
i hope they have "ephemeral" messages early on in the game. not gonna be fun to have to deal with admin only messages that would be visible to the whole server
"Don't touch the fish"
Hopefully some sort of sending a message only to a specific player
Would be interesting if chat is separate from the console or if it's combined or maybe there is a separate UI for it
Well, it likely wouldn't be hard to implement as a plugin if its not built in I am sure there would be one somebody would make day one anyway. 😄 (Even if crude)
Depends, I guess if they allow servers go add their own UI then ye
Otherwise it wouldn't work well
@prisma agate check out slikey’s newest reply, he said that because clients and servers don’t share a codebase, it “has interpolation but prediction is currently limited”
OH you probably already saw lmao
I haven't I don't use Twitter
Ye so it just interpolates the last known velocity, like Minecraft does, and then I assume it just rolls back if the desync is too large
no it was quoted in here up top.
I thought this meant that it interpolated between its position and the next server position, using last known velocity would be prediction no? Even if rather primitive prediction
Is it possible to send live texture updates to player in this game? Anyone knows?
It is, hot reloading is possible so you can change any assets live
Is hot reloading going to be for all or just code?
All!
In case you didn't know, there's going to be a live visual editor to change things real time
What visual editor are you talking about exactly? I mean you’ll be able to change things real time if you’re in the server code on intellij or doing models on blockbench but there’s no first party visual editor app
If u do it with just positions it means u gotta use the them to predict the next, bcz the server didn't send u the next one yet, so that'd be prediction, and a simple way to do it is u take the difference of the 2 prev positions, which estimates the real velocity
it'd be way simpler to just tell the client the exact velocity (even if there is a delay between updating it
https://hytale com/news/2025/11/hytale-modding-strategy-and-status
That way at least u r not estimating, u get the exact velocity, just delayed
Oh I completely forgot about that one, we literally got a video on editing vfx 😅
To kyren: yeah I guess so, position is a thing that is pretty independent of game content so predicting position with velocity is probably safe for them to have on the client
Actually here would be the client. The client predicts while the server delay processes. - Essentially makes the client responsive to you but that delay will still exist.
it'd be way simpler to just tell the client the exact velocity (even if there is a delay between updating it
^ This would make the game terrible to play because you would press W and two seconds later your moving... - While simpler would give such a bad experience it would kill any game.
In reality you are "two people" you have a "server you" and a "client you" - You only see the "client you", every one else sees the "server you" (You request your "server you" to move, while that happens "client you" is already moving) - Two different entities essentially. Client Prediction and Reconciliation. You can use interpolation to make things very smooth and natural.
Unless your comment is to an entire different system that I couldn't pickup. 😛
Ye it was about a different thing but I now realize it makes no sense
If the server computes velocities the delay will be horrible, so I guess the client must be aware of all the kinds of movements that r possible, so it can compute velocities on its own
Yep! - Essentially you simulate just enough of the game to make it "locally responsive" everything else is on the server.
Ye although I now wonder how does set velocity works in mc
Well velocity is just part of the simulated compute, both the server and client know this. So the simulation plays and if the server and client didn't "agree" on the numbers the client gets reconsiled to what the server states (rubber banding effect)
Bcz personally I haven't noticed a 200ms delay between input to activate an ability and the time the velocity effects me, but might've been bcz of localhost
I mean for plugins with custom abilities like dash, so only the server initiates a player.setVelocity(), the client can't predict it bcz it's not in the base game
There must be at least a one way trip delay from the server to the client to notify it "hey ur velocity is now this"
Well, this is where I find this architecture amusing. Just because the server is the one with the plugins doesn't mean the client doesn't get the code. - Roblox uses this methodology, same as Fortnite.
I think if code plugins who manipulate/invent player mechanics that are not "manipulations" of already created mechanics (such as Minecraft event interoupts) will be a thing the client does get.
as a full time dev in a different industry this was both super complicated sounding and also a "oh that makes sense tho". well written
Thanks! - I try to explain it the same way I learned and understood it. (Definitely not super simple stuff until it clicks) 😄 - God I can't spell
yeah make sense, at the end of the day as a player i want what i see to look and feel good. everyone else seeing me slightly couple coordinates at a diff position wouldnt matter to me. vs trying to reconcile and sync server and client calc
Ye, which is what tazer and I discussed, some kind of command system to allow the client to compute things that weren't hardcoded
Or alternatively a sandboxed PL like Roblox does with lua
is it fair for me to assume that the default port for servers will be 25565?
No. Not fair to assume likely to think though
Probablh not, to not conflict with mc
ohhh yeah
Only solution I could think of, (Or just pull a Minecraft and don't allow custom inventive systems) but rather force us to develop entire systems within the existing ones. - Basically either as free as Roblox (using their methods) or as free as Minecraft (using their methods) 😄 Unfortunately nothing has been stated that I have found about what they are doing... 🙁
Anyone tryna cook up a factions type of server?
Have server requirements been mentioned yet?
not yet
keep forgetting no links are allowd here, even the built in discord gifs
Stickers are allowed!
it's hard to tell probably, because requirements can drastically change in a month, if they do some optimizations.. I would bet that it's less resource heavy than optimized papermc server
Sure…
Could be true in some circumstances probably, like in minigames
I'd be extremely surprised if it were lighter in a normal survival setting though 💀
We know a couple things, based on what they revealed about client requirements:
Saved game size (in Exploration mode):
~661 MB for 5000x5000 blocks
~27 KB per 32x32 chunk
Mods will increase requirements. Large content packs, complex server logic, or extended view distances can push both the CPU and the GPU harder than in the baseline game.
RAM:
Singleplayer, with dedicated graphics: 8 GB
Singleplayer, with integrated graphics: 12 GB
Multiplayer-only: 8 GB
CPU: Intel Core i5-10400 (or equivalent), AMD Ryzen 5 3600 (or equivalent)
After that, focus on a solid multi-core CPU and enough RAM rather than a top-end GPU.
So basically, we know the world sizes. We also know the obvious: mods will increase server load. And we know that the RAM utilization is negligible* with the base game (based on the difference between singleplayer and multiplayer). We also know the game prefers multi-core processors, and runs fine on a 6-core 4GHz chip (when running integrated server and client).
Negligible is probably a generous term to use, but we can at least say that base game isn't going to be using more than 8GB lol
So we have an upper bounds to work with at least
And we know these are system requirements, which gives a bit of wiggle room for the OS and the likes
regardless just stash up on powerful ryzens
if it's ACTUALLY multithreaded, powerful single core cpus are still powerful since you can utilize the whole thing. We know that the dimensions can load in parallel, however, we don't know whether an entire world can be run in parallel, which is what matters. On 2b2t, if a region is rendered by a core and someone builds a lag machine it only lags that sector
LOL we know it'll be multi-threaded but I don't think we need to stock-pile EPYCs or anything 🙈
Maybe it's not a safe assumption, but I'd imagine they're multi-threading the NPCs, since that would presumably be the laggiest part of the game
Just get a powerful Ryzen, focus on making good content, keeping people and getting people to your server. It doesn't matter. Renting a good server is dirt cheap anyways if you can actually monetize a hundred CCU
CCU?
crazy cool users
CCU almost never gets used for Minecraft but ever since Roblox blew up it's a more commonly used term, it's at least what I use
Yeah, I found it was pretty common in specific places, but I had never seen this before
Anyways, the answer is simple, look around for some black friday sales, find a cheap Ryzen server, if you gain too many players you should have more than enough money to literally rent a better box
OVH still has a black friday sale going
Yeah, the specs seem like they'll be pretty affordable at least for small scales
Everything got more expensive but renting servers got MUCH cheaper
We'll see how it goes with their 60 player tests though, and release
You can literally rent an 8GB of RAM, 4vCPU VPS for $4.20/month now anywhere in the world... Shockingly servers got cheaper
Wtf??
Lol if I didn't have my own I guess I'd be renting
A server for 24 months or a stick of DDR4?
Did the cpus get worse
No. The software did. No one can optimize anything.
funny thing is my bf found an online server hosting that told him it was the price of 0, for 24gb of ram and 4 cores
Oracle's free tier, I know what you're talking about
yeah he told me shortly after
Oracle constantly goes out of a stock and you risk getting randomly banned by Oracle
still somewhat surprising because of the ram prices
well now thats something i've not heard of oracle's free tier before
I'm paying $4/month for 512mb ram on digital ocean, what provider is this
Don't bother with it, someone in this server said they constantly go out of stock
OVH and they have servers in North America, Europe, and Asia. If you're in Europe, go with Hetzner
Hetzner and Netcup are actually what you want to use if you're European, they are known for IDing people due to so many bad people trying to rent their cheap stuff, but usually won't ID you at all if you're from Germany
See usually I could argue this given it were the client side of a game with the graphics and all, but your absolutely correct! - Servers are failing to be optimized. 😂
I LITERALLY witnessed single shard servers with hundreds of players back then, even one with 200-300 players with 50-100 mods loaded in back over a decade ago.
im not gonna, i was putting in a comment of mine here. since i had a very recent tidbit from the other guy
(i probably couldn't host a server properly rn anyways)
OVH is not perfect but, what matters is you can grab cheap dedis and that's where the actual value is since, with a VPS you always risk getting throttled since it's on a shared machine
As long as you upgrade your account to the Pay as you go tier, but only use what they offer in their Free tier, you can get free tier servers almost immediately on request. Source, i use Oracle free tier. They aren't the most powerful obvi, but I pregenerate chunks and usually only have 4-5 people on at a time and they work great even for modpacks
Hetzner you can get a bonkers dedicated machine for about $120
Latest Ryzen 9 - I forget their whole setup, but its like 198 Gigs, and terabytes of storage. Not to mention a 40Tb bandwidth cap a month or so? Not bad AT ALL. OVH I hear is close to the same just a different structure.
Racknerd charges $10.60 per year for a 1GB of RAM, 25GB of storage VPS as well... However Racknerd does not let you delete your account even when you're done using them
i used to use OVH back in like 2017-18 are they still viable
The people who had their servers burnt down however aren't too happy
they had a datacenter literally in cargo boxes (not a dexter reference) and the wood floor caught fire and tons of data got destroyed
i did not know that wow
BuyVM also has dirt cheap VMs but they got bought out and they are NEVER in stock. 512mb for $24/year
OVH has two listings for their bandwidth one says 10mbps the other 400mbps
In Asia the bandwidth is more expensive
I can't imagine there's going to be OAuth for accounts released at launch, but it would be great if there was. Does anyone know if we will have a way to verify account/username ownership from a web api?
Probably not at launch but it should be out quickly
Just take a nice look at OVH, make sure you dont use a VPN when buying, use actual info, as many of these other server hosts have tons of antifraud measures and are paranoid due to tons of people trying to do bad stuff with them
As I can see OVH only has a north american server available in Canada
OVH has a pretty good BF sale, picked up a dedi with an additional 32gb of ram that was offered for free 🙂
I can imagine prices will also go up, once micron officially leaves the consumer space
Uh... They are allocating their consumer RAM production to enterprise, meaning they will be supplying more ram to the datacenters. That means all the RAM they would have made for consumers go to those datacenters that let you rent it out
I have no idea what the percentage was but they should have just put consumer stuff on pause while focusing on this AI boom instead of closing the Crucial brand completely
keep in mind it's actually really really really difficult to build factories that make the chips and all these computer parts
hmm, but isn't the ram being supplied to AI data centers very different than traditional ram, a company say like OVH would typically use for their dedis?
So when the crypto mining boom happened before the AI boom, the government and the enterprises that used GPUs already weren't affected since there is a conveyer belt/production line already promised to those industries, meaning no shortages
& apparently OVH will see a price increase of 5-10% due to the RAM issue
keep in mind building a factory that produces RAM isn't as easy as it's sounds, it's not like Costco can make Kirkland brand DDR5 sticks out of thin air, even if you had $100M in liquid on you, it might takes YEARS for you to build a factory that can produce those things, and you'd need the people 😭
Yeah that will suck
Even the manufacturers who know how to build these factories would take at least two years to get one up and running
It also amazes me how a company like micron can just leave? overnight ~20% of the competition is gone lol
What programming langauge will be used for plugins? sorry I forgot.
Java, you can always check the blog post on it lol
Java, but you could port the server software in any language you want as long as it follows the game's protocol
I'm an automation engineer in real life, one of the bigger projects I've done was well over 500M in equipment alone not even including engineering hours programming the system to work. If you're making something like computer parts that price immediately increases 2.5x or more
You have no idea what you’re talking about.
Please stop
Lol you should've seen what he said earlier and then deleted
I'm confused. What is he wrong about?
Nothing 💀
Just a simple supply and demand effect
I mean, technically they aren't allocating their consumer RAM production to datacenters that you can rent out. They're allocating it to HBM modules for AI datacenters, which are not really being rented out
Do y’all think Hytale will release documentation for modding prior to the games early access launch?
I think if they're able to then they absolutely will, I just hope they aren't on too much of a time crunch now that they've given a hard and fast release date and have to release it the same day as launch
I'm sure they'll release a ton of info leading up to the launch tho, like so much more than we're getting now
I hope they would, but not a guarantee.
makes sense why you can't just build more factories
bureaucracy
I guess people tend to underestimate how much goes into building these factories, even if you have the knowledge.
If it was easy and a safe bet, Micron would just build another factory and continue to sell both AI chips and desktop/server RAM. But they don't.
Just spin up another RAMBuilderFactoryFactory.java
Maybe we can stick AI into Hytale in places where it shouldn't be. Instead of a world generator, let's instead use Generative AI and prompt it to generate a world for us. Much easier right? That way, Hytale can benefit from all the AI-enabled hardware that everyone seems to want to mass produce and hoard these days. /s
Every entity that moves in the world also is just being goverened by GenAI prompts. Stuff like "there is a player 2 blocks in front of you, what do you do?" or something.
Oh wait I forgot, Hytale is not owned by a publicly traded company that has to answer to AI hungry shareholders. Well there goes that idea...
The truth is: generative AI WILL be used in stuff like world generators but not in the way people think
people will use ChatGPT/Claude Code to write the code for the LINEAR functions to randomly generate worlds, a neural network might design and be used to improve upon already existing algorithms.
Ah yeah, I meant jokingly more like using GenAI in the running game to generate worlds via prompts. I also use AI tools during development, where they make sense.
Yeah I know you were joking 😂 but scarily enough people don't realize that AI is now being used in everything, even if a neural network isn't deciding or generating it, a neural network will likely have had involvement in writing the code for it
the real people who actually will figure out how to make money with AI without burning precious VC money are the people who figure out how to make specialized AIs that aren't resource intensive. Ones that can actually fit and run locally on a phone
Step 1 of that VC process is going pretty well: get everyone hooked on the product and establish a monopoly.
I swapped from Evernote to Notion this week. I used to pay $3-$6/month for Evernote. It's now $25/month. So I am cancelling
Ah that is step 2 in full swing: squeeze all the money out of it
AI actually is useful but the problem is it's like Mr Meeseeks from Rick and Morty. He warned his family NOT to use the Meeseeks box for anything other than basic physical labor
yeah I was loafing around. I don't think even Notch alone could even fix the RAM shortage with his money, that's how badly expensive it is
Amazing idea, perhaps we could also use AI to play for us, surprise us! We just sit back and relax...
I think there's hardly a monopoly, if you mean with large language models
The locally available programming models are already capable enough, genuinely
Genie is out of the bottle in that regard
Maybe not a monopoly in the models, but definitely with the hardware
Maybe for the training compute, but actually running inference isn't all that expensive
the problem is now crucial pulled out of giving consumers good RAM so now more shortages, higher prices
AI is popping just like the "metaverse"
AI is actually being used, the metaverse wasn't being used
I can't even remember what exactly metaverse was
ready player one
The Zucc's odd take on reality
GPUs released 5 years ago can still run modern AI models quite effectively, the shortages might not effect companies actually looking to setup a temporary programming LLM
Beast movie, wish we could have a game like that irl BUT metaverse wasn't anything like that...
It was just a VR social media platform basically, with the typical ad-space and data collection
the book apparently is really really awful like it just sucks
Have you read it?
Have you heard of Unsloth AI? How long do you think it'd take to train a free model on the docs
I heard they are making a second movie
I haven't been I've seen some reviews saying it's bad
Training takes absolute ages unfortunately
But the open models are good enough, and feasible to finetune with something like unsloth
Well you should judge by yourself honestly but still it doesn't matter anyways
it's like spacejam 2 where it's a reference fest
and that's what it really is like apparently
I liked the movie
When I had a 3060, I was experimenting with coding LLMs and it was genuinely plenty capable with just 12GB of VRAM
A 3090 has double that, and we've got better CPU implementations now
Not to mention things like bitnet and the dedicated neural processors some have been trying to bring to the table
Intel basically has an AI consumer card for sale as we speak, last I checked
Is LLMs going to be powered by nuclear?
With any luck lol, better that than coal
in like decades when the government finally approves a plant
Kinda crazy 80% of the emissions in 2024 were from 4 AI companies
But but! Hiroshima mannn and Chernobyl
Is that actually a real statistic? Because that's fckin ridiculous
Yeah
lol the government is afraid of nuclear radiation but not polluting the ocean to the point where there is tons of plastic in our fish
A google away will even tell you that AI companies are draining rivers
The places where there are cooler climates, they can actually just funnel the heat outside from I heard
Don't forget to use a paper straw though >3
uhhhh i dont think thats true
That's how Hetzner I think actually cools their datacenters
The government is just made up of people voted in by other people
Nobody knows how to assess the danger of things they don't understand 😍
Honestly what's true or false nowadays, you don't know
well 80% is bs lol, that doesnt even add up in the slightest
80% of emissions sounds like a lot, what kind of emissions? The ones produced by electricity production? Because even so, have you seen the stuff coming out of things like steel foundries
I heard some want to build datacenters below the cities to warm them up during the winter
That actually might be a good idea
In the US energy report for 2023, it was about 30% of server electricity usage
They predicted a doubling for 2024
its some baseless fact he made up lol
This idea is actually being used already in some places. In The Netherlands, there are companies like LeafCloud that do exactly that: instead of one centralized datacenter, they deploy small server racks in places like care homes and whatnot to produce the heating for those places. These then group up to form one cloud platform.
Kinda wish we could use nuclear 😭
What are you guys planning on cooking once hytale modding releases?
Some pasta, most likely
Minigames and some QoL stuff if I end up hosting a survival server with friends
I would love to work on an immersive RP server
I might first just chuck a server somewhere to play on it with some people in private. To experience the game in its current state and just enjoy playing it. After that, might start to explore some ideas.
Since they will add vc, it will be pog
No offense intended but do you actually know much about the things you talk about here? Why is the server-plugins channel your home? LOL
Because I have no friends and I like to talk to people
i honestly doubt that after the claim that data centers cause 80% of emissions lmao
I was just curious because you didn't know about mixins, and have some very strong views lol
I would probably call myself the creative director with basic coding knowledge that gets thing done.
interesting!
Most of my modding experience is fixing or customizing other people's mods so 🙈
I'm definitely amateur lol
Haven't made basically anything from the ground up, outside of some server utilities
And those definitely weren't game mods
I tried mc modding but it was too complex for me
When I first jumped into modding Minecraft, I didn't know we could see the remapped source 🙈
But I was still doing code injection, and just doing it blindly rofl
Reverse engineering things that didn't need to reverse engineered, because watching a tutorial is just beyond my comprehension apparently
I tend to stick with plugin development. A pretty mature ecosystem these days. Can sometimes be challenging (in a fun way) to push whatever Vanilla mechanics there are to their limit to create custom stuff.
ah but you learn the most that way
Yeah I don't really do content mods, though mine is kind of the opposite
I worked exclusively client-side, still limited to vanilla mechanics
Yeah plugin development is sick. I kinda stopped building complex stuff when I learnt about the new features like custom uis, shaders. Too hard
That's why I am so passionate about hytale's modding scene because it lets you be creative fast without having to know workarounds... just a simple library that got everything you need
I definitely learnt quite a bit, and once I found out I could just read the source basically
I was like "holy cow this is so easy" LMAO
That'll be when it's mature though, presumably
Up until then we'll be using incomplete APIs and code injection (the mixins)
Yeah I haven't made custom UIs or shaders or anything either
I've made some chest UIs I suppose lol
Yeah chest uis is easy, you just swap the textures. I am talking about these custom characters that you format them to a correct position etc
I hope we get the same api they used in the trailer
It seems usable, unless it was made purely for showcase
yeah I still can't wrap my head around custom font stuff in minecraft
hytale's custom ui possibilities are really exciting
I tried the font stuff a few times, found it pretty hacky. Maybe there are better libraries for it these days
It is also ridiculous how we have had to resort to changing fonts to make new GUIs in Minecraft though. Minecraft did finally start to add custom UI things, but super limited.
The fact that it took Mojang more than a decade to even give basic UI support, which is still clunky
A simple api for aligning stuff left middle right would be cool
I have a lot more faith in Hytale when that's the bar LOL
If they expose things like screen size and the likes, you can just calculate those kinds of things easily
Honestly, the lack of custom UIs has been one of the biggest blockers for proper RPGs in Minecraft. Everything has to be done in clunky ways that kind of hinder the experience.
Literally, even if you want custom blocks. You have to retexture the current ones. Best block is the noteblock because you have like 50 different noteblock types (for now)
Genuinely, everything in that game is a work-around and clunky
Considering minecraft is making a good chunk of money with bedrock, they sure don't make the experience easy 😂
The whole mechanic to get a datapack correctly turned on for a client is also clunky
I mean, look at that one big RPG server, they literally have to put you in limbo until everything is enabled. Great user experience.
Protocol hacks and things Mojang never intended
Client-side entities, even with resourcepacks and everything were ridiculous
We had to put custom models on sticks and force them into the offhand to render static HUD elements LOL
Yeah, Wynncraft was actually groundbreaking at release though
Back in 1.8. If you wanted a title bar, you had to spawn a wither. Make it invisible and ultra small and follow the players mouse
That's true actually, hilariously so
Minecraft has made big leaps though, with the introduction of display entities. But still not the ultimate fix to all problems
And how many resource and datapack versions do Mojang rotate through each minor release?
Alot
Have you seen Cymaera's work with them?
Breaking changes every snapshot
Such a stable API amirite
You can barely call that 😭 Most of it it's just replacing the game with a resourcepack
They royally fcked semantic versioning so hard that they literally gave up
They just changed all the version numbers, and didn't even introduce parity
And that was after toying with datapack versioning and adding minor revisions, because whoever was working on that didn't realize you weren't supposed to push up the number every time
I doubt the UI was an intended feature. Some high iq probably found a way to put a texture
It's the intended way for us to create user content, and it's not even properly supported
Considering hytale is working with the 2018 version, they are ahead of their time!
In a few years time, 1.21.568 will finally drop that introduces vertical slabs
The fact that they hired Gnembon and didn't implement a real scripting language is even more absurd
Genuinely who could look at mcfunctions and think that was a good idea
Maybe the reasoning is that since people make stuff with it, they shouldn't improve it
I think you mean 29.90.568, because they just changed them all lol
It's not like they listen to creators, hypixel was screaming for a transfer packet since 2014. They added it super recently
They listened though! Only a decade later
Can't wait for the server blog 😩
For me, peak was when the Mojang devs answered a question about vertical slabs with "it is not aligned with our vision of the game". Like hello, we are talking about a vertical slab, not about introducing a sphere.
Maybe we can convince Microsoft to buy Hytale too, and get the sort of community interaction we deserve
You shall enjoy 8 new blocks!
Here's copper. It is used for lightning rods, and that's about it. It can oxidize though!
They're just afraid to break the golden goose
So naturally they do literally nothing ever, and ignore everyone who tends to the golden goose
If you think about what people were doing in Minecraft in 2014
Makes you wonder why Microsoft even bought it
Now it would be a mistake to make tons of breaking changes every version to the point where everyone struggles to keep up, I can understand that. I also think that the amount of legacy they had to deal with carefully was also holding back development a lot. I am happy to see that in the past two years, they seemed to have worked through lots of their tech debt and are now capable of introducing stuff like customly scaled entities, custom gravity, display entities, a start on custom UIs, etc.
Minecraft 1.10 mentioned!!!
1.10 was actually neat though, wasn't that the colours one?
The llamas were cute, and they added more falling blocks
Except in 2025 it's
8 blocks
78 breaking changes
No 1.12 was colours
The thing I dislike the most is the half-baked features. Like they would introduce things that are done 100 times better in mods, but then fail to properly execute on it. A good example are hoppers.
1.10 added 4 blocks (magma, bone, nether wart, red nether brick) and 3 mobs (husks, strays, polar bears)
Did servers go back to Java as their language of code or are they C# / C++
Oh it was the biome mobs
They're Java again
Don't forget the new command, /teleport
Yeah servers are using a java api
Updates like 1.13 and 1.16 were pretty neat, with finally big changes
Honestly I am not complaining about minecraft not adding blocks and stuff. But at least if you are making millions (billions perhaps). Allow people to change your game effortlessly
The fact that we have to think really hard to name updates that were actually interesting is its own problem
Best part about all of those updates are the fact that they broke every user-created mod on the planet too
so I'm guessing the issues similar to MC servers struggling with high player counts will be present on hytale servers or iam wrong in that statement?
We don't know yet, but that issue isn't due to the language
Hytale is commited to optimise their modding scene. So they will do their best to not have issue like minecraft's
All I know is that in c#, gurun (the backend admin responsible for hypixel pocket edition) was able to host 500 players in 1 server.
and I assume their backend is c#?
What's more relevant is that Hytale has experience with developing a server with a high number of concurrent players
They likely have considered ways to handle numbers as high as their own
Yeah, they've also said that we will be able to turn off features to optimise our servers
Definitely not all on one server, but even if it can manage double the player counts, it'll go leaps and bounds
That's kind of implicit though, and isn't necessarily about optimization
Yeah everything is speculation right now
Hmm fairs
I wouldn't say everything is speculation 🙈
true
True but they said that can also boost performance since its not loading what you disable
With the considerations they have taken and their statements I wouldnt expect more or about 50 concurrent. It could be more, it could be less. But voxel games are by nature computationally expensive, this is my estimate from my personal experience with developing my own voxel engines. I wouldn't be suprised with this number, nor would I be suprised if you had conditions to allow 250 players. - It is all up.in the air, but their view distance is what makes me wonder. (World gen, view distance. Etc)
These specific instances: Higher counts for limited worlds likely purely mini games, controlled worlds, no worlds gen, and limited entities, etc
Full world sim I would expect more than Minecraft, at least not anything major, maybe give or take 10-20 more players. This is where I can see 50 maybe a few more, but nothing by any leaps and bounds.
Is world gen infinite? How does zoning work
Well we know that we have per world tick system at least. Unsure about the world gen.
I'm surprised you're expecting full sim to fit more than Minecraft, we'll definitely see
Though I'm curious what numbers you'd give Minecraft when it comes to "full sim"
A lot of how we make large player counts work is by simply not ticking anything lol
Or are you thinking exclusively vanilla behaviour, so everything is ticked and always
There are infinite worlds
They will be disabled at launch because zones are strange with them, allegedly
They were talking about their tick system, and ECS these two things very easily cater to the performance of entities. If they followed through with that I wouldnt be suprised if it were a bit more performance even with their view distance increase, especially after the world is generated.
What would you consider the capacity of a Minecraft server?
Is hytale using single core
no
Without going bonkers on machine and price and server software (vanilla) 10-20 seems to be its good zone.
You cannoptimize for more via Paper, pre gen, hardware, etc. But 10-20 is my range.
That's what I was thinking too
I actually can't think of a single time I've ever used a vanilla server for more than 5 people
So I don't have any personal metrics for how badly it gets out of hand, but I'd expect it to choke atrociously quick with the single thread lol
I've also pre-generated every world I've ever hosted rofl
I've always set world borders at like 4k radius
Which is all just band-aids for the fact that Minecraft just instantly melts down with world gen stacked on top
Completely unplayable if you have the server utilization limited
Well if Hytale was developing for the server experience I would love to see them support high cpu core counts, 32, 64, 128. Not your normal 4, 6, and 8.
I think it would also be intriguing to see them spin up threads for each player/entity! Though we quickly enter enterprise specialized hosting now to the realm of Eve Online (and we all know NOBODY will spend another 20nyears just for their networking)... 😄
This above is just me wanting, nothing is based off of their statements, but it would be nice (though not "plugin" friendly.)
Although I was thinking, and worlds to me sound like instances, do they plan to have an orchestration server which manages "worlds" because well that would be built in routing for instances and quite nice given each world has its own tick system.
Built a Hypixel like network with dispersion built in.
I'm dreaming. Sorry! Don't take my words as fact what so ever, still speculation and dreams (likely to be broken)
You might have to create your own orchestration
More realistic ^ 😄
Confident you won't have to create world load workarounds like bungee. Just a simple transfer packet
See this is what is very nice. Should be simple and seemingly made for scale huh.
Exactly
Is there currently any information communicated about hardware requirements for hosting a server?
Nothing stated explicitly by the team, but the CPU and RAM requirements should be in the ballpark of the recommended specs blogpost. Most likely beefier CPU than the recommended, but RAM allocation will probably vary based on the player count/mod count
Hi is it possible to develop scripts on hytale yet? Is it known what programming language does it use for scripting?
Well surely you'll be able to turn a lot off I hope because the specs for users involve both client+server. So servers oughta be lower, which would be nice as recommended client specs would be very high for a server.
Java
Hytale uses JAVA for their modding API. But they are working on a visual scripting editor too
cough use kotlin
holy cow it will have java plugins
nahhh start from html guys
HTML easy
Hold my Scratch
slikey
The server manual is coming in later than expected as we are still tweaking some performance issues on high player counts. Last test with 56 players showed us some systems misbehaving and we had to discard the data. I hope that tomorrow we have ironed most things out and get a more stable idea.
What do you guys think?
Hey my estimate above was fairly accurate. About 50!
Thats what I think.
I wanna know what's real reason behind performance issues
Isn't that tweet just a way of saying they don't know what the real reason is either
Since the test was invalidated and they have to try again
Excited for the results though
I'm curious what the specs were for the 50 player test
At least 128GB of RAM 💸
Jokes aside, curious to see the specs when there are a lot of people on it. Of course in singleplayer it also runs the server, but that data isn't saying much as you're the only one "online"
Yeah
i know the game is rough, but 56 players is... rough.. theres already youtubers with pretty large followings telling their communities theyll be making their own servers so this is a none starter. wonder what system is bottle necking the network layer.
When in doubt add more RAM until it fixes itself
Maybe there will be a Folia or a Multipaper or ShreddedPaper coming to Hytale if it can't handle many players on it's own
My computer when I already have 16 exabytes of RAM (the maximum amount of RAM before 64-bit memory addresses run out): 
That one gif of Kylo Ren yelling more
Tbh server jar systems perhaps are gunna be a thing just like mc maybe?
It's presumably CPU limited
I still think the CPU will need to be beefier because it'll be handling multiple clients instead of just the one, but yeah, RAM shouldn't need to be 16 GB in smaller servers, probably closer to 4-8 allocated
Hytale is already multi threaded, so it would have to be a more novel solution
Or I suppose a way simpler one, assuming it's in need of optimization
"Multi-Threaded" is sort of a magical term. There are many was to make any application multi-threaded, and some of them are indeed quite novel
Hytale's server will be shared source, so if people have performance improvements, they can contribute them instead of having to start a spin-off project like Paper
The idea is that if they're struggling already, then you need even more novel than the current implementation
Right, that's the beauty of the way the hytale devs have described the server. So novel high player count servers can be created without having to consider them break-offs like paper etc.
We'll still likely see a branch of sorts, if legal
I'd imagine they'd only accept low-compromise optimizations, though I suppose it could just be configurable
The lengths that Minecraft server owners go to maintain playability are quite brutal when it comes to compromises
Slikey already said it's perfectly acceptable to abandon their server code and create something completely new. It's part of their paradigm
Well yeah, there are no legal problems with that
The legal problems would come in with sharing modified versions of their code, depending on the license of course
they didn't say 56 was the limit, just that some systems misbehaved at that point, it might have used 2GB ram and 20% of average CPU..
We already know from minecraft that some people want servers that can host 200-500 players. Some have even achieved more. With the popularity that we expect from Hytale, I'm sure we will see people wanting big servers. I will be really interested in what the official server implementation can handle.
Yeah, just happened to be the size of the test, and big enough to highlight issues and invalidate the test
We'll see how it can actually do soon hopefully
Id geniunly love a massive survival world with 200-500 people tbh
I'd be surprised if it weren't possible with the right specs
im guessing the right specs for that would be 32-64gb of ram
LOL I think the CPU would be the more expensive part to shop for in those circumstances
i always forget about CPUs tbh (im used to ram being the main deal lol)
Ram will be the easy part, it the number of cores that are both available and more importantly that can be fully utilized
They put a decent amount of emphasis on how performance was worse in singleplayer, and that CPU mattered more
Their recommended specs were a 6-core 4GHz CPU if I recall
We talked about the requirements a bit here previously
mines 8 core but with a avarage speed of 3 Ghz
I just say 4GHz because that's the peak/turbo of the chips they listed
You could definitely run singleplayer at least
Just as a reference point, Folia servers for minecraft recommend 16 cores minimum, but that is for 200-1000 players
Holy cow, and that's just with region-based multi-threading?
but fewer cores are of course ok for fewer players
I just looked up Folia like 10 minutes ago after it was mentioned lol
Those are very impressive numbers
im kinda curius how expensive new gaming pcs are (my pc is 5 years old)
They have a blog post on the hardware requirements if you're concerned
The game should be relatively easy to run, especially when you're connecting to someone else's server
We have identified that the server went out of control and spawned 3000 NPCs which were all active... So yeah.. It wasn't the amount of players that caused the issues, it was that our spawning for NPCs malfunctioned :'D
~Slikey
Oh!
3000 active entities that probs can crash any system lol
I wonder how many entities you'll typically get per player lol
Like the 70 hostiles we get in Minecraft
3000 would put you at just 40 players
i somehow at one point when i was working on a server i gliched spawning to the point i had 800 enties in a 10 x 10 chunk area ._.
Where did he end up saying this?
Entity control/limiting is one of the techniques used to help servers in minecraft.
reddit(dot)com/r/hytale/comments/1pgpapl/comment/nsuls9w/
Yeah, and we go pretty far with it
That's kind of what I had in mind when talking about compromises
Skipping ticking entirely a majority of the time, wiping out items, quicker despawns, smaller caps, etc
just curius how long do yall think basic community plugins will take to form once hytale comes out?
im saying equvelents of minecrafts essencials x / multiverse core if hytale supports it / luckperms?
I wouldn't expect Essentials to take any time at all basically
LuckPerms might not be very necessary, I don't think they've talked about permissions like that at all
Thanks
I think luckperms will be ported to Hytale, and I will work on some open-source plugin for land protection, but I would appreciate if more people would help, it will be in Kotlin, I think I could utilize their area selection tool from creative mode for area selection - instead of chunks, because they will be 32x32
I'm mostly curious about a equivalent of multiverse core now if its even possible tbh.
Land protection will be nice
They might already have a permissions system in place that's sufficient
And multiverse might not be needed at all either, like how it's not needed in Minecraft anymore
towny hytale edition lol
They do happen to be directly catering towards the towny and cozy players
In Hytale, if you're a cozy player who enjoys farming, building a cute home, caring for animals, cooking, and simply wants to escape the real world, we’ve got you covered. We’re now assembling a dedicated team focused entirely on bringing this vibe to life and creating a living world filled with warmth, charm, and good vibes. ❤️
[image of an interior featuring bookcases, table, chairs, and other misc. decor]
That will be great if you will make a claim system. I hope you could also offer a chunk by chunk option too as that is much lighter on system resources. I know the chunks are large, but it still is a useful option in many cases.
I like the Residence plugin for area protection, it's very popular in Czechia/Slovakia, but I don't know much about other plugins.. Residence is very advanced, you can do subareas and you have flags for everything, per player or globally.. what is common on some survival servers? Some chunk claim system?
i just realized how small memory wise hytale is to mc when it comes to world storage
The ≈27 KB per 32x32 chunk?
I actually thought that their specs of 27kb per chunk was a typo, I can't believe it would be so small. I'll have to see it to beleive it, lol
Sounds good, better for performance, but it's often negligible, you can also make cache of what claims are in specific chunk and when someone breaks a block, you'll check only like 1-4 claims if block is in their area, you don't have to loop thousands of server claims to do math.
I don't actually know how big a 2x2 of Minecraft chunks is
Do either of you? I could just check real quick lol
~661 MB for 5000x5000 blocks
i mean there needs to be 2 typos then
hmm i work with world painter on mc and my 6000 by 3000 maps take up 1GB
1,397,124 spread across 253 region files
Region files are 32x32, and a Hytale chunk is a 2x2
So you divide by 256 to get the 21.57KB?
My Cobblemon server comes in at 34.56KB per Hytale chunk
Unless the thinking above is wrong
I have a superflat with an average of 2.1KB, but it seems to be due to outliers
Median was around 5.03KB
There are a few outliers in all of them, so the averages might not be a good choice lol
Presumably has to do with the fact that the regions aren't necessarily populated
whats a claim system
It's where you take ownership of land, preventing others from damaging it
Like "claiming" your home would prevent griefing from others
ah so like minecraft world guard thing
Yes, exactly
But for the players to use, as opposed to admins
16x16 block chunks on one of my minecraft servers range from 8 to 12 kb, so 32x32 block chunks would be 32 to 48 kb equivalent. So maybe their stated 27 kb is not unrealistic.
How did you measure the chunks? Or were you also toying with the region files
Yea, I just took region file size and calculated
My math from before is definitely flawed in that a region file can have anywhere between 1 chunk and 1024, so I should probably test with a fully generated world
Yea, I took regions that I knew were full
Very nice
Their claims are very realistic, attests to their honesty
Just in case we were worried about it all being an ultimate ruse LOL
should have some poll of who plans to develop in java vs kotlin to get some sample size based off this community. worry the ecosystem will become segmented between the two
It's definitely not going to be a problem if it's segregated
I plan on doing pure Java to begin with, haven't worked with Kotlin before
But I was actually looking into it specifically for this, as I also haven't built something from the ground-up in Java before
And I heard Kotlin is nicer for that
there should be no poll, just dont use java, it's simple as that
It's not like the Java mods are going to be incompatible with everything else
using java is like using sticks and stones instead of real tools when repairing your car.. and there is no excuse you havent worked with kotlin before - it literally takes few hours to catch the syntax differences
What exactly is so much better? Is there something more efficient about the bytecode Kotlin spits out?
it's like night a day, not worth explaining or arguing.. with java you have to write ton of code to accomplish the same.. you don't have nullability or auto casting.. and every time you need to specify type of variable even if it's obvious from what you are putting into it
Nulls and implicit casts sound virtually inconsequential 🙈
It seems strange to have such a strong opinion on it, if those are the highlights
It's not a strong opinion, Java is legacy language for which is too late to fix how it was designed, Kotlin was designed to fix all java mistakes and it replaced java, it's just improved syntax you can use, you'll compile it to JVM bytecode anyways
you can look up some examples, check for instance this:
github(dot)com/rixafy/NetteHelpers/blob/main/src/main/kotlin/org/nette/helpers/completion/ControlCompletionContributor.kt#L28
and it's not even that big of a deal, but I don't want to imagine how that 1 line of code would be written in Java.. you would have to store and check if parameters.getOriginalPosition() is null, then if not you would have store parent variable, then check it's instance, if it's not what you are looking for then return the code, if it is.. create a new variable to hold it.. just suffering
They're nice to have. Not as essential as some people think. Java has numerous serious issues (as do many other popular languages) but it's overhated at the moment, possibly as a result of certain programming YouTube influencers who I will not be naming.
Boilerplate is annoying for smaller projects but less significant as a proportion of total lines written for larger ones. I wouldn't prototype in Java, there are languages built to make that as easy as possible.
Kotlin doesn't even fix some really glaring problems, like lack of true primitive generics. That's a JVM issue. If you need efficient data structures for primitives you will have to reach for libraries like fastutil. Otherwise, Kotlin has nice ergonomics and reads better than Java. But I would expect no significant different in developer productivity for large, well-organized projects. Your language choice will not be your bottleneck, just use whatever your team is most familiar with.
I think I'll write my plugins in 50% Java 50% Kotlin
Maybe in time I can get some Clojure in the mix
Balanced, as all things should be. 😆
Personally I'm not a fan of mixed codebases, but if it works for your team, it works.
In all seriousness Kotlin is very nice developer experience, but if you haven't kept up with Java, it's pretty neat how far it's come. Project Valhalla is such an interesting case study in implementation research and trade-offs and when it finally it makes it into a release it's going to be very cool.
I can't wait for Project Valhalla. It's been a year away for the past 10 years!
Java is not a legacy language. Maybe legacy versions of java are legacy but the programming language itself is not legacy.
If you think old languages are legacy know it's newer than python
I've always seen Kotlin as an experiment used to test new features before slowly bringing over to Java. The language being slow to change is a good thing in my opinion
are we really comparing language from 1996 with language designed by JetBrains developers from 2016? yes we are
everyone can have different preferences, Kotlin is objectively better on many levels
some people like Java even though have they seen Kotlin, there is nothing wrong with that
Kotlin is a better language than Java. But the difference between the two is not as pronounced as many say. There's some nuance there.
Ultimately language choice isn't a very big deal. Use what you already know.
I don't think anyone is comparing anything to 1996 Java
Java 25 came out 3 months ago
well I used Java from 2013 to 2018 for almost every day, and Kotlin onwards
makes no difference, you can't just make nullable/not-nullable types in java.. best you can do is to annotate nulls
Optionals?
Java is a type of language where in order to sort a list, you need to write code block with Comparator etc. in kotlin you do list.sort { player.name }..
makes no difference
Java 25 is VERY different to for example Java 8. It does make a huge difference even if a single feature you mention isn't there
I know it makes difference, but there are hard limits, that's why Kotlin was invented
Collections.sort(names)
Same thing in modern Java
I'd prefer to use Collections.sort(list) if your element type implements Comparable- oops @sacred tulip beat me to it 😆
this is simply not true.. discord even wont allow me to post it, but it's way longer, and you still need Comparator
ask GPT java sort List of Player (player have getName and sort it by that) and it shows you the real code
ask GPT
gotta be ragebaiting
and even ask for java 25 version
I am looking at the docs for Java SE 8. There is a public static method sort in java.util.Collections that accepts a List<T> where T extends Comparable<? super T>. I have used it, numerous times, it it does not require a comparator, only that your type implement Comparable. Is that what you're talking about?
you still dont get it, my example was list.sort { player.name }, it sorts by Player::getName()
it's list of Player object, not string objects
Implement Comparable for your Player class...
zzzz
Not a perfect solution since maybe comparing by name isn't the canonical way of organizing players.
everything you can do in kotlin, you can do in java, but with additional lines of code or way longer lines, I'm not arguing with that.. some people may like Java, I know lot of them
and im not going to be using AI or google every time I want sort a list when I can just create a .kt file that has way more intuitive syntax you can learn in about an hour
Sorting lists is uncommon relative to how often it appears on e.g. coding tutorials and interview questions. And I can't say I've ever had trouble remembering how to do it in Java, on the rare occasion it's necessary.
But yes. It is more verbose.
You don't even need a block if you use some helper methods on Comparator, IIRC.
I was using sorting as an example, there is no single thing java can do better
It's easier to find examples and tutorials for how to do things in Java
Java works better than Kotlin if your team is more used to Java. But in a general sense, you are correct. It's just that it's not that big of a deal.
Real talk, I would much rather write Rust than either Java or Kotlin. 😆
Well but Rust is from different world, if you know Java, it takes you few hours to understand Kotlin.. but a few days or weeks to understand Rust (how it works, what can it do..), Kotlin is just better syntax for JVM ecosystem, that's all..
JetBrains employee levels of fangirling
Actually yes this is also a factor, but that depends on the team, it isn't really a fault of the language per se.
If you paste Java code into Kotlin file in Intellij Idea, it automatiaclly do conversion to Kotlin, it's not much different
Yeah, true, they're not comparable when we're talking about Hytale. You can't (feasibly) use Rust for Hytale dev unless you want to do some fun stuff that will likely cause Hytale's developers to look at you funny.
so every java tutorial is kotlin compatible, but yes, Java has more content online
Exception handling is better in Java than Kotlin
True, Kotlin doesn't have a native multi-catch yet
I'm looking forward to when the Hytale devs give us more detailed info on the server so we can talk about that, lol
I'm afraid it will be after the release
Because until then they can change API so much that if they said something yesterday it's no longer true
I would at least like more info on server requirements so I can prepare
For how many players are you preparing?
I would like a little table that shows for example 10, 50, 100 etc players
Why would you prepare to make a server before playing the game what if it sucks
So are you guys planning on hosting your own content, or just making a plugin and letting other people run it if they want?
Because there's a 5-10 year development promise and the source will be shared
IIRC Slikey said they did perf tests with 53 players (no idea under what conditions) and they ran into some problems. More information to come tomorrow?
I posted more info here, he said it was because server randomly spawned 3000 NPCs
We have identified that the server went out of control and spawned 3000 NPCs which were all active... So yeah.. It wasn't the amount of players that caused the issues, it was that our spawning for NPCs malfunctioned :'D
Oh. 😆 that would do it.
I'm simply more interested in programming and game development than I am in playing the survival sandbox game
The server manual is coming in later than expected as we are still tweaking some performance issues on high player counts. Last test with 56 players showed us some systems misbehaving and we had to discard the data. I hope that tomorrow we have ironed most things out and get a more stable idea.
Server manual was expected 1-2 months after release anyways, at the release date we will have to decompile it and look how it works, or give it to the AI overlords to generate docs for us
If I were AI and saw whole source code, I would create pretty usable docs
At least the server will be unobfuscated.
At the moment, I'm most interested in building bespoke minigames, which means hosted. I think that will be the best way to experiment and learn the ins and outs of what's possible to handshake and with the protocol. But I'd also really like to get situated in the community in a way I haven't in MC so community contribution I think will be a good chance for that, too.
I'm planning on making a few minigames and survival tweaks
If they end up being any good, I plan to share the repos
You just feed AI with whole source code so it has context and then you ask how player join event works.. and AI figures out there is no such thing but there is PlayerConnectEvent so it prints you the details
Girl, you can't trust it to do that
Especially not all in one batch
Other option is to search for that event.. depends what's faster.. but you can let your IDE index the source code and some AI agent will have summary of what is where
Minigame focused for both of you, I see. Fun stuff. That's what I did in my Minecraft development days. I was only ever server-side.
Yeah, same
I've done exclusively server-side, and exclusively client-side later
But never any content mods intended for both
You can get pretty far with quality-of-life mods with just client-side stuff, so that's what I switched to when playing more survival
It’s not good with high context. AI does well with any content at the end and start. But loses memory in the middle. I’ve used pretty much every mainstream model on 200k+ line code bases and it’s absolutely useless.
I had big plans for a massive client (NeoForge) mod, and then Hytale's re-acquisition by Simon hit me like a truck. And now I'm here.
I don't know how long you've followed map creators like Simon, but way back we had all sorts of maps
Some of the most amazing ones were done by FVDisco
And the possibilities with Hytale kind of reminds me of that again lol
Very nostalgic seeing the Hypixel name and the potential for minigames, shooters, adventures
I saw some cute models in #1440422085750030457 with like, Plants vs Zombies assets
And we used to have all sorts of games like that too
I joined Hypixel a bit late for the adventure map train. But I did actually go back and play through Herobrine's Mansion on my localhost a few years ago. It was quite impressive considering the time period it was made in.
I'm interested to see what they're gonna come up with for Hypixel on Hytale. I don't know about you, but (Minecraft) Hypixel has felt stale for a while now. Hopefully with Hytale, and full creative control given that it's their own game, we'll see some really neat creative stuff.
And that's not even mentioning what the rest of the community will do!
Yeah, I haven't looked at Hypixel's server in a very very long time
I remember a few games though, like Creeper Attack
And I vaguely remember Bed Wars taking off
If you wanted to look at some retro experiences
Some classics are Creeper Attack, Cake Defense, Blocks vs Zombies
If you search on YouTube for Minecraft maps and use filters like before:2015-1-1, you can get a little glimpse back in time lol
It's disappointing how corporate and deaf Mojang eventually became
I'm a Creeper Attack enjoyer, though these days it's probably impossible to get a game to start. Most of the games I used to like are pretty much dead. The server is 90% Skyblock now.
Mojang is disappointing. It's like they don't even understand why people liked their game so much in the first place, and they're just adding new stuff for the sake of it.
It's genuinely hilarious that Microsoft bought an indie game that revolved around small communities and user creation
And then disrupted the ecosystem with a new EULA, killed Minecon, ignored user feedback, avoided innovation
And only now 10 years later, they're finally inching towards actually supporting user content and making things data-driven
And even then they've been doing such a poor job at not breaking packs/mods that they just ditched their versioning scheme entirely 😍
And that's just keeping the focus entirely on Java edition
The primary version of the game is on a whole other level of cash-grabbing and poor performance
I'm one of those people that thinks Minecraft lost its sense of direction a long time ago, and never quite recovered from the addition of being able to sleep through the night LOL
So the base game has been struggling since ever, and modding has been struggling since ever as well
Promises of a modding API floated around for years and years and years, all while the game remained obfuscated and they were completely silent
The most popular mod for Minecraft was a mod that vaguely improved performance and added a few extra resourcepack features
Literally the most basic extension of user-created resourcepacks too, and it was wildly popular
I think it is kind of shameful that the base game is so bad to the point where many people need sodium to even run it
The reason why the vanilla+ movement is so popular for example is because the base game is so lopsided towards environmental additions but even those are poorly executed
I think they don't realize just how important modders and content creators have been for their success. Even if most people don't mess with mods, they've probably watched a YouTuber who does.
The vanilla+ movement is literally only a thing because isn’t feasible to extensively mod post-1.12 minecraft to like gtnh levels without needing a ryzen 18 1218050272849x
They're weirdly reluctant to hand over what they see as "control" to their community, even though their community is probably more than half the reason the game sold as many copies as it did.
I hope that the servers wont be 20 TPS, that would literally kill me
They're just such cowards, they're even reluctant to develop on their own LOL
They promised a rework of the 1.9 combat and can't even follow-through on that
So scared to break the golden goose, that they're just starving it
What tick rate are you hoping for?
I mean it doesnt have to be counterstrike high at 128 but I would want it so vehicles, animations and particles are smoothe enough to actually enjoy.
It just feels hard to fathom that a giant bloated game is technically the same thing as what I played on an iphone 5 in 2013
and by particles, I don't mean client side, I mean positioning on the server
I'd imagine all of those things will play just fine on the client and be interpolated when needed
I thought maybe you were hoping for something like 60 for competitive gamemodes
It could be customizable. Though that would be challenging for them to implement.
It's probably static and it's probably around 20
I feel like they would've mentioned it when talking about networking if it varied in resolution significantly
Yeah doing some sort of predictability/interpolation would be amazing
But who knows! We'll hopefully hear about it in some server performance blog post
I wonder why hytale chose to have java for server and not c#
They went back to the old engine, which was Java
They're pushing for early access as soon as possible basically
Ecosystem (library support), familiarity (they're targeting a community that is probably pretty familiar with Java)?
They probably talk about it somewhere.
I guess cross compatibility also?
Especially since curseforge is in the spotlight for modding
They're probably quite familiar with Java themselves as well
They were Minecraft modders after all lol
I guess it makes sense if the client is in java is to make it consistent.
The client is C# from what I heard, just the server is Java
Modding and plugin ecosystem is def better yeah
I missread engine thinking about the client whoops
Also, fairly easy to make a cross-platform application without having to compile for different architectures. The same .jar should run on both Linux and Windows. Which it needs to, because most people will be hosting on Linux devices, but playing singleplayer on Windows.
But in 2025 it's on par with java no?
Modern compilers? You'd still need a decent amount of platform specific code, as opposed to targeting the JVM
but there is now a Single unified runtime no? the SDK runs everywhere..?
I don't think I understand what you mean
I haven't written much cross-platform C#, so I can't really comment.
I'm fairly sure the client application Hytale is distributing is native bytecode, not CIL or whatever you'd normally compile C# to.
Should read into Unified .NET
Oh I see what you're talking about
Yeah that is true bytecode sandboxing I guess is difficult with c#?
I think they're just using Java now because they were using it then, the last time around before cancellation
And they probably picked Java back then due to their familiarity, and possibly their audience of Minecraft mod developers
Correct, they're using ahead-of-time compilation to ship native code for the client. This is what makes multiplatform a bit more difficult
This is just my opinion though, we could just ask at some point probably LOL
They should rewrite the client in Javascript for true cross-platform capabilities. /s
oh hell naw
hytale on webasm when
client should be rust compiled to webasm for maximum speed and compatibility
If I can't run Hytale in my webbrowser I'm gonna riot...
unfortunately for you, riot sold it :P
WebAssembly hytale
I want the true Minecraft classic experience
A web JVM client fully loaded into RAM and illegally distributed across the world wide web
0.0.23a_01 is apparently still accessible on classic.minecraft.net LOL
Makes sense, I can still get all those old versions thru Prismlauncher and presumably the vanilla launcher too.
I hope plugin hot-loading is a lot more stable
Then again, not every plugin is written by java devs...
I'm actually super curious about the live-updating of assets
Like, how live we talking? Lol
I didn't actually see the source but people have been saying it's planned
It'd be really convenient if they've got a filesystem watcher for assets. So you can just save whatever you're working on and it'll reload in-game automatically.
Yeah would be intersting to see how their runtime asset manager will function
Probably a hybrid of manually requesting assets and a background process for fetching live asset updates?
I don't think it will actually be "live"?
This is probably how it's gonna work. Automatic updates for some things, manual for others, hopefully most won't require a full restart.
If it's live-live, you could even push assets dynamically generated
Or assets from separate web servers, assets uploaded by directly players