CraftAU shutting down.

    Joined
    Apr 25, 2017
    Messages
    6
    Reaction score
    0
    I've worked with a few game companies, have consulted for many more and have many friends in the industry, I've never seen it work like that :) The base of the game is quite solid, we've always maintained it. In almost all game companies, especially the larger ones, feature development is first, bugs and performance overhauls usually come in late alpha and beta. Sure, you need a solid foundation first, which is what we worked on for the first few years of SM development, as well as the many years of work to create the in-house engine we use. You'll find our recent update isn't a rework/redesign of the engine or overhaul of the game, it's simply massive bug cleaning. The fact that we've scheduled most bugfixes/cleanup for late alpha/beta isn't news, we've been very transparent about this from the very beginning. I can't think of any game I've seen/been involved with that's done otherwise.

    I've seen internal workings of AAA companies, mid-tiers and indies. I know of and am quite close to people who work in all spectrums of the industry. In fact, Schine is part of a couple of game studio programs. If you work (or have worked) in the industry, I'll be happy to listen to how you've seen it work, I just have never seen it work as you've described. I suggest contacting me via PM or Skype (I believe you have my details already), send your credentials, and we can discuss the inner workings of game companies. There's always something to learn.
    I have a ton of friends in the industry. Koderz and I both live in Raleigh NC and know quite a few people at Epic Games and RedStorm here. I also know people at basically every other well known studio (CCP/Bungie/Blizzard/Bohemia etc ) and also others near the industry like ILM/Autodesk and companies you likely use their products in your work like GitHub/Atlassian. I currently work in the industry, so does Koderz.

    Now I never said you should do all of it as you go along, only that you don't put it all off to the end. As for having a solid core, I would have to agree/disagree with that. When we played and ran RS, there were a litany of core networking issues, collision issues (pretty sure these still mostly exist), and don't get me started on the file storage and database. I also never said you were completely redesigning the internals of the engine, I was giving an example.

    If the companies you've worked for haven't done bugs as they move along, and at minimum kept an eye on obvious performance problems then they're companies I wouldn't want to work for. Code built on a bug is inherently bugged so when you come back later and fix the underlying bug you run the risk of completely breaking the parts on top of it so waiting until the end for all bugs is a horrible idea. Obviously there's a big push for bugs and late performance improvements but they need to be handled to a degree all throughout the process, which is what we had a problem with. Continually adding features and degrading stability/performance, especially to near unplayable levels, actually ends up slowing you down as well as annoying the players when it's in an open state like this.

    I don't have your Skype, Koderz might.
     

    DukeofRealms

    Count Duku
    Joined
    Sep 4, 2013
    Messages
    1,476
    Reaction score
    1,616
    • Schine
    Now I never said you should do all of it as you go along, only that you don't put it all off to the end. As for having a solid core, I would have to agree/disagree with that. When we played and ran RS, there were a litany of core networking issues, collision issues (pretty sure these still mostly exist), and don't get me started on the file storage and database. I also never said you were completely redesigning the internals of the engine, I was giving an example.
    I'm sure if you reported those back when you were running that server, we still have them documented. We document all issues reported to us, particularly from server admins.

    I'm not going to go into detail here on how our development works and what I've seen, it's not really relevant to the thread here. I believe schema and other members of our dev team have discussed that with either you or your friend in the past. Feel free to contact me via PM.

    You can do that here: https://starmadedock.net/conversations/add?to=DukeofRealms
     
    Joined
    Apr 25, 2017
    Messages
    6
    Reaction score
    0
    I'm sure if you reported those back when you were running that server, we still have them documented. We document all issues reported to us, particularly from server admins.

    I'm not going to go into detail here on how our development works and what I've seen, it's not really relevant to the thread here. I believe schema and other members of our dev team have discussed that with either you or your friend in the past. Feel free to contact me via PM.

    I think relevancy went out the window a while back since this thread started for a closing server and then derailed long before I got here.

    You may have them documented, but nothing's been done about them so that doesn't really mean much. That's like me writing a reminder to myself to do something but never doing it, what's the point? Ignoring the development practices, we're talking about glaring, MASSIVE, problems that at this point have existed for 4+ years. That's what I was originally asking was why they've persisted this long and honestly why it still appears nothing is being done about them. There's no reason a game server for a game like this should take hundreds of GB of disk space for <1000 people There's no reason the same server should take 500+mb/s network upband to support 50 people. The collision has a ton of room for improvement as well, granted that's a harder thing to do. Those are what I would consider core systems.

    No offense, but if I thought it would actually amount to something I would contact you and explain the list of issues that plagued us back then, most of which appear to still be here, but honestly I've been down that road now a couple times and I've lost track of how many times Koderz has attempted to no avail. We stick around hoping SM will actually get somewhere but honestly are losing hope that it ever will. :(
     

    DukeofRealms

    Count Duku
    Joined
    Sep 4, 2013
    Messages
    1,476
    Reaction score
    1,616
    • Schine
    You have my contact details, AndyP , schema and I have gone over this on more than one occasion with you (it's unclear which account you really are, so I'm not sure what we've specifically said to you). We work with plenty of other servers and communities at the moment, who provide us invaluable information, which we've used in this update.

    I can give you responses to issues you list, and if they're valid and not scheduled to be fixed, we can look at that. I'm more than willing to help you regain access to your account, I'm more than willing to bring up previous conversations and take a look at those. A conversation on here is going to go nowhere.

    If you don't think it's going to help, that's fine, in that case, I'm not sure what we can do to help you.
     

    Matt_Bradock

    The Shrink
    Joined
    Aug 4, 2013
    Messages
    798
    Reaction score
    464
    • Purchased!
    • Thinking Positive
    • Legacy Citizen 5
    As CoolDev2 mentioned, it is possible to write a voxel engine in C++ that handles also a physical simulation in a relatively short time.

    I feel like the problem is, that Java has its limits. It's great for a great many things, but rendering dynamic 3D objects and then calculating collision simulations between them is actually pushing the very language to its limits (that and being so prone to memory leaks that plagued earlier Starmade and still does happen every now and then).
    It is a great achievement to write a game with these capabilities in Java. But not efficient. There is a reason even Minecraft is being ported to C++ (if hasn't been already) and other space voxel game devs chose different runtime environments and languages for their games (looking at Space Engineers for example, but I could also mention Empyrion, although their physics simulation is practically non-existant)

    Looks like to me, schema's greatest skill - his experties in Java - is also his greatest bane. Starting the whole thing in C++ or Unity would have allowed for much cleaner work, allowing more focus on features and stability and less on working around the limitations of the engine and runtime environment to make it still work somehow.
     
    Joined
    Apr 25, 2017
    Messages
    6
    Reaction score
    0
    As CoolDev2 mentioned, it is possible to write a voxel engine in C++ that handles also a physical simulation in a relatively short time.

    I feel like the problem is, that Java has its limits. It's great for a great many things, but rendering dynamic 3D objects and then calculating collision simulations between them is actually pushing the very language to its limits (that and being so prone to memory leaks that plagued earlier Starmade and still does happen every now and then).
    It is a great achievement to write a game with these capabilities in Java. But not efficient. There is a reason even Minecraft is being ported to C++ (if hasn't been already) and other space voxel game devs chose different runtime environments and languages for their games (looking at Space Engineers for example, but I could also mention Empyrion, although their physics simulation is practically non-existant)

    Looks like to me, schema's greatest skill - his experties in Java - is also his greatest bane. Starting the whole thing in C++ or Unity would have allowed for much cleaner work, allowing more focus on features and stability and less on working around the limitations of the engine and runtime environment to make it still work somehow.
    Java does have it's limits, but those problems I listed above aside from collision aren't affected by the language. Collision is to a degree, but honestly it can still be faster within Java. Also it's all to easy to offload some things to c while keeping most of it in java (pretty sure they already did this with noise)

    Java also isn't prone to memory leaks at all. it's a precise garbage collected environment so it's impossible to actually "leak" memory within the confines of the language (assuming the language runtime doesn't have bugs itself) The only thing along this line that you can do in java is not remove your references to something you don't need as if you have a reference the GC won't delete it. Same goes for C#.

    Now for Unity. While the core engine is C++, the environment you work in is C#, which is subject to most of the same performance characteristics of Java.

    Now for Space Engineers... Guess what? It's C#! GitHub - KeenSoftwareHouse/SpaceEngineers

    Empyrion I believe is Unity which means C# as well. (Or technically it means Mono, but it could be any of the .net supported languages but they'd all have basically the same performance as C#)

    I do agree that working on an existing engine would be better but Unity wasn't as well known when Schema started as it is now. UE4 wasn't publicly available until 2014. CryEngine... yeah nvm...


    Again, this isn't to diminish what Schema has done, but it's not yet a language limitation. I just wish the team would have put effort into areas other than just features before now.
     
    Joined
    Oct 10, 2014
    Messages
    6
    Reaction score
    2
    • Legacy Citizen 2
    Oh man, that sucks. I'm going to miss having an Aussie server. Hell, that was one of the things I loved about Starmade in general- these indie games rarely have a large, dedicated Aussie server.

    *EDIT*
    Didn't realise this had become a code discussion thread. I'll just take my very limited, working knowledge of the subject and leave, I think...