Development started in 2012, with the hanger release coming out August 2013... So that's a long 4 years and 5 months in alpha. Which is insane, what will beta be? 3 years?
******Official Star Citizen / Squadron 42 Thread******
- Thread starter Raumarik
- Start date
Development started in 2012, with the hanger release coming out August 2013... So that's a long 4 years and 5 months in alpha. Which is insane, what will beta be? 3 years?
Permabanned
- Joined
- 3 Jul 2008
- Posts
- 3,762
- Location
- My fabulous ship
Just looking at the comments - you need to understand, 3.0 was a complete rewrite of what they did for 2.6. If you haven't read the patch notes, I suggest you do - but you'll be there for a while.
No content has been added so far - everything you see is the initial beginning for new functionality - miles ekhart is the first npc mission giver, npcs are in the first iteration, landing on planets is new, exploreable planets are new, object container streaming (only part in), interacting with objects and taking into your ship, interaction feature (f key), ship wrecks.... this is just off the top of my head.
All of this is just the first iteration which will be expanded on. Now that it's in the game, the game has a massive task of optimization, the frame rates you speak of (20fps) is because of the servers redraw rate, not because of your PC. I have been on servers and had 60fps no problem, this was because I was on my own, the bugs in the game currently gradually bring the server down to a 20fps crawl on the client - and in the beginning of 3.0, servers were still going down regularly, this is an alpha and a new build.
by the end of this year, don't expect massive amounts of content - mining will be in 3.1 and by 3.4 (December) we should have an initial glimpse of what we will soon enjoy.
No content has been added so far - everything you see is the initial beginning for new functionality - miles ekhart is the first npc mission giver, npcs are in the first iteration, landing on planets is new, exploreable planets are new, object container streaming (only part in), interacting with objects and taking into your ship, interaction feature (f key), ship wrecks.... this is just off the top of my head.
All of this is just the first iteration which will be expanded on. Now that it's in the game, the game has a massive task of optimization, the frame rates you speak of (20fps) is because of the servers redraw rate, not because of your PC. I have been on servers and had 60fps no problem, this was because I was on my own, the bugs in the game currently gradually bring the server down to a 20fps crawl on the client - and in the beginning of 3.0, servers were still going down regularly, this is an alpha and a new build.
by the end of this year, don't expect massive amounts of content - mining will be in 3.1 and by 3.4 (December) we should have an initial glimpse of what we will soon enjoy.
Last edited:
Not really, I'd honestly say until 2.6.3 there wasn't really an alpha. What we had before was a fancy version of space invaders at best (Arena commander), a tech demo racing track (still is), a ship model showcase (hangar model) and a half baked but somehow still cool FPS module.
One of their biggest weaknesses has been having to redo parts.
FPS module, now done twice. We even sure it's not going to be redone again?
Ship models, many of them have been done at least twice. Sometimes because of basic errors like making the Mustang with the wrong size reference character so the cockpit was physically too small. It's since been remade again, third time lucky?
Hangar module - we're now on version 2 of it with two maybe 3 of the hangers having been made twice for the "room module" which was being worked on in 2014.. but we don't have.
It's pretty late but I'm sure there's been more redone from scratch..
I am still absolutely gobsmacked that the client FPS is dependent on the server FPS. It stinks of some extremely broken fundamentals in the engine.
But it isn't all that insane. There are hundreds of games that have 7+ years development cycles. I woudl honestly say we might have the Beta in 3 years and a year long Beta.
But you really need to consider when you compare general development time frames.
- Most are already established studios with established procedures, not starting with 4 people and trying to hire the other 400 odd people needed to develop the game as they go. Not trying to get all the procedures for a medium sized company in place, working through trying to flash hire 400+ staff, trying to set up the studios across 3 continents and get game design docs available for all to work on the same page when you don't even know what your doing with the game yet due to the nature of open funding and development.
- Funding explode in 2013-2014 (great! But a double edged sword). This comes with a huge amount of fallout where suddenly the team had to decide what to do with the funds. Do they keep to the same goal and pocket the monies? Or do they expand the goals, increase the game scope early on and extend the development. They put it to vote. Early backers were excited and voted for more content.
- To boot it is actually two high end games not one. Regardless of if it uses the single client to access they are clearly very much two separate games in terms of what and how the player interacts with the game
But Beta may be in 3 years and a year long. This isn't steam early access where for the most part games are 60% there before we even see them or the "alpha/beta" that companies like EA do with public, those games are 90% there at that point. They are very different. If you was to see what the games were like at the start you would see that the first 3 years is almost entirely different tech demos, background tech, engine manipulation, lore and setting the goals for the next 5 years (ish) of development.
Take SW:TOR (Star wars: The old republic) a game based on a universe with one of the richest laws written and all the design assets already having all the concept art they need and yet they still took 2 years to write the lore prior to other development starting. That puts the game around 6 years in development for an already massive IP with a team already in place with probably the worlds largest games publisher behind it and you are expecting a start up company trying to complete two games of which ones game scope goes far and beyond what the above had certainly at release to be all done and finished in 5 years?
This is the problem with open development and giving access from day 1 of development. People generally don't realise the time frame or see other projects that are generally much much later in development and expected this to be the same. Tbh CR himself set it up to fail in terms of time frame in 2012 tbh. That I think came down more to creating excitement and getting the funding at time. Even with the smaller scope it would have been a minimum 4 years development cycle.
But at same time for things that are in kick starter campaigns it is worth always reading up and being realistic to yourself. Someone selling and marketing will always be ambitious and embellish the truth because their job is to sell their vision to you to try and make it happen.
Half of the service for the client run server side. Longer term it will offload the client and increase performance. Short to medium term it will be an issue till server meshing and the container streaming system is in place. This is something that was always going to need to be developed alongside what we have with 3.0. The amount of data passing through them is huge and because all data passes through it is a roadblock. Animations, spawning in a couple of large ships causing hitching to all in the server. When UDP packets get dropped or delayed that can wreak havoc
If it was closed development that wouldn't be the case as they would not have followed the same development pipeline and thus these services would be in before the public ever saw it. But people really need to look at the general public as buy in QA team for them. It is literally all about testing the game mechanics as they are added, testing server performance, client side performance and so on.
With that I am generally getting 20fps+ and when interdiction is off 30fps+ which at this stage of development is absolutely what would be expected. Nothing stinks of things, you like to pick holes in things without even looking at what is going on. The engine is technically broken, that is what they are working on (the data culling via the container streaming, culling and server meshing as noted above) and have been for 2 years approx. It is a system they are writing from scratch to make this work. It is looking to be at earliest September to see that. Then expect another year or two to actually get it doing what it is meant too.
Squad actually has (had maybe now, haven't played in a while) the same issue that as soon as a server fills up with players performance drops significantly. All development for online multiplayer games go through this stage. And they are also on 50 server pop and getting 10fps reports compared to 50fps with only 10 people in.
@ curlyriff ~ i appreciate all the above points, but i still feel we're ending up at the same old debate - bloating. the original KS concept was perfect. adding all the other stuff in may very well enhance the game and take it to the next level, but i still maintain that a lot of what's going on now is simply now what fired our imaginations on the KS. Adding a pointless [to me] racing aspect is taking away a lot of creative talent from the KS concept. adding all this language cac is taking creative talent from the KS concept. adding sandworms is too. all this stuff could have been drip-fed in later. if all the talent was focused on what was originally touted then we would probably be further down the road - how far is probably incalculable, but my point is that it is really frustrating [to me] to have been pimped a next-level Wing Commander, then continuously getting told that it's way overdue from the dates they announced at the start, and the reasons for that are because my next-gen WC is now becoming WC-and-Forza-and-GTA-and-whatever-else.
but y'know, this is all deja-vu at this point, your post suggesting things are still may 4 years off at least is pretty depressing ~ for those who haven't fallen to apathy anyway.
but y'know, this is all deja-vu at this point, your post suggesting things are still may 4 years off at least is pretty depressing ~ for those who haven't fallen to apathy anyway.
Fundamentally, the client game thread should not be waiting for network updates, it should be happily chugging along at whatever FPS the client is capable of, with the poor server FPS manifesting as janky movement of replicated actors (or not, depending on how good their interpolation and other trickery is).
They are what, 5 years in now? That this is going to be an MMO isn't feature creep, it's been 100% the plan from the start, that they haven't even addressed the very fundamentals of an engine to support that should be setting off massive alarm bells.
Caporegime
- Joined
- 7 Apr 2008
- Posts
- 25,115
- Location
- Lorville - Hurston
seriously day by day i feel this will fail as a project.
Sad and kind of wished EA or another publisher just bought the rights . Say what you like about EA etc but they at least get a game released thats playable
Sad and kind of wished EA or another publisher just bought the rights . Say what you like about EA etc but they at least get a game released thats playable
As the scope has expanded it's possible it's not going to be the game you, or any of us, want it to be. I would rather be hopeful a PC only game that is trying new things will be good instead of seeing some people hoping that it will fail for some bizarre reason.
EA don't have a very good history when it comes to taking over games and companies.
Instead of playing the same small scope games publishers rehash year after year, waiting for something new with features that have never been done before doesn't sound so bad to me.
EA ? what? they don't improve anything they buy.. heck I'd rather it just not get made than end up like Sims in space with 20+ DLCS.
You're right. Better to pay for all the DLCs and in game items before the game launches ;D
Except you don't have too. It is all optional which is nothing like EA and the worst Sims game produced to date (although the actual build tools are good).
But the point is that everyone is receiving all the data of all other 49 players as if they are directly looking at them right now. The system currently isn't culling any data so although we cannot see it the game things they are as such. There is no culling of information at all right now. That is the fundamental issue. It is not directly the server itself, it is the connection between client & server and they have been working on this for 2 years but needed the fundamental base engine and game mechanics in to start this process. They are addressing it. That is how long things take. The complexity for these things that have not been done or tackled by say a larger studio, Bethesda, Activision, BioWare, Bungie, Rockstar North for instance haven't done it previous is time and monies. They are waiting on a company like CIG to do the work and then follow on once the dev's have the experience that they can hire later. Welcome to the industry.
You cannot start to do data culling on multiple physics grids if you have not yet created the physics grids to determine the data. You need to realise the first two years are almost a right off. They were base tech demos in house to see where the engine was and they started creating their first round of base tools to complete their initial ship pipeline and animation system with motion capture.
Then it expanded in 2014 to allow for fully realised procedural generated planets and cities. This data that also has to be there before they can create the culling system to suit that. No one suggested feature creep to why the network hasn't been finished but the fundamentals is actually to try and get the engine to run 50 to 100 person servers without the culling system by optimising what is old CryEngine code (that lumberyard also still use) and that takes years of iteration. They have almost got there with them hoping to get that nailed down to constant 30fps for everyone and stable by end of March.
It is also the reason that for them to get the amount of data at the detail they want into one instance is waiting for the server stuff touted to be arriving in September before expanding the Stanton system to the 4 planets and similar.
I know a fair bit about the industry thanks, all this jargon doesn't wash with me. Network LODs and culling are not some kind of mystic voodoo that CiG have invented....
They should not be struggling with the fundamentals this far down the line. They should have a solid, stable, performant foundation working to build out the game on. All the evidence points to them being several years away from the stage where they can sensibly begin building out all the rest of the game.
Urghhh I am not just chucking jargon or suggesting they invented the principles but your being purposefully agnostic in the process of thought.
Of course LOD's and Culling exist. That still doesn't mean they can create them without having the data to do so which would mean they required a lot of the base tech in place. For instance how can you write code to cull information about a procedural generated planet city where it in itself is new tech if you don't know what or how to do so yet.
They need to make sure they create the right elements with the right number of container streaming systems built in so the culling system works with it. And sorry but this stuff really would take at least a few years to do from scratch. We are on about finishing up the foundation to build on at the moment. That is what this whole year is about.
If they just done it for a single planetary system then a lot of what they are doing wouldn't be needed but they are working on making it scale to the point that the only reason we are getting long term (in premise) 120+ known systems is the time frame to actually produce them. In theoretical terms if they get these foundations right with it working in connection with the systems being discussed then you could theoretically have unlimited systems to explore.
So again we are back to the point that up to 2014 then it was base premise of small portions with a much smaller development team (less than 150 people). They started working on the network stuff around 2015. This is a reasonable time frame to write code from scratch and make sure it works for a game/engine that has never scaled to such extremes. The building blocks are there and some are not even in the current build. You keep making out it should be 6 months and bam it is done and networking would all be sorted.
That just isn't the case. The networking and what they are doing to separate operations into dedicated server systems such as the overarching economic system are significant deals. They have the interdiction system on another isolated server. These are all just being brought online at moment to offload the central server system and allow them to do dynamic updates so when the main servers spin down and reboot they can pick up the hotfix without the client side doing anything.
There are things that have been done on all other games but it is the combination that makes it complex and the time frame that your being unrealistic with.
Edit: And with things like this they are having to try and introduce an equivalent to static precomputed PVS in a none static state. That hasn't been done as far as I am aware as it is conceptually extremely difficult to do. As you said culling has always been there but the type of culling and the level of culling that can be specified is also important.
From what I understand it is the expansion of the base CryEngine (anti)portal system that they are first introducing. At the moment we don't even have that their in terms of client to server data. Yes it is visually culling otherwise we would see through into ships for instance but the data is still flowing from client to server and back to all other 49 clients.
Last edited:
I'm not denying that what they are doing is complicated, all game dev is. I am generally very sympathetic towards gamedev difficulties because it's my day job and I know that pain (as it happens I've been working on NPC state replication the past few days).
I'm also an (ex) backer of SC and want it to succeed, but it looks to me like the project is not on the right track.
CR may be a good creative director, but SC is *desperately* in need of an experienced producer who is given equal power to CR in order to make sure a game actually ships.
It's patently obvious CR is getting carried away with his spaceship fantasies, with bigger and more complex ships appearing all the time....without actually knowing if they will *ever* be able to get them working in the MMO environment they want to build....because they have neglected the fundamentals for far too long.
I'm also an (ex) backer of SC and want it to succeed, but it looks to me like the project is not on the right track.
CR may be a good creative director, but SC is *desperately* in need of an experienced producer who is given equal power to CR in order to make sure a game actually ships.
It's patently obvious CR is getting carried away with his spaceship fantasies, with bigger and more complex ships appearing all the time....without actually knowing if they will *ever* be able to get them working in the MMO environment they want to build....because they have neglected the fundamentals for far too long.
isn't it possible for a company like RSI to get a pre-written utility to do this, at least as a jumping off point if nowt else, or are all the existing MMO games using their own proprietary code? in much the way things like GNU seem to be able to produce a virtually exact copy of Microsoft Excel etc, i'd have thought a big company that's going to depend on MMO elements for a massive part of its product would be alble to do similar, either reverse engineer an existing thing or take/buy an existing solution and tweak it as appropriate? given they were using CryEngine and now Lumberyard, is there no similar working games [ie, MMO needing a lot of data for a lot of players] that use the same thing that RSI could use?
does this mean that if one of those element servers goes down the whole game is buggered? i understand the concept that if, eg, the American servers went down then America couldn't play, but if one server holds all economic data and that goes down, everyone will not be able to do anything involving economies?