The game is never - ever - getting made. The vindication only comes in the form that I was right since I called them out in July 2015 and said that by over-scoping the game, they killed the project.
Right now, there are bunch of lunatics parked on Reddit saying that because 3.0 has barren moons - aka planetary tech - that I've been proven wrong. Completely ignoring the fact that my statements have been precise, to the point, and not open to any misinterpretation. That being, the game as pitched, cannot be made. It's got nothing to do with any one thing (engine, networking, physics, AI, rendering, terrain etc); it's to do with ALL the things that we as developers put together to call a GAME. And this shit here, is what they simply can't make.
First post, so first let me thank you on sharing your insights so far.
As non developer might i ask you, while you mention those interconnectivity, to share your thoughts on the whole cutting back on network traffic by reducing the connectivity between people within and outside of objects?
They've so often shown off characters waving at each other while being in different objects, and have so many hatches instead of actual airlocks in their ship design, that it just screams to me that they've got to revisit those designs and decisions, which would be another letdown for their shrinking fanbase. Also questioning how you think they'll actually tackle boarding actions during space battles when they're now speaking about heavily instancing stuff.
Personally i think Roberts is so far doing the most expansive brainstorming session in history. As i've seldomly heard/read them discussing any implications for the gameplay systems of their decisions so far, but tons of nebulous general ideas of what systems are going to be in the game.
As someone who has built networking tech over the years, it's a very difficult subject to explain to non-programmers. The long and short of it is that it all goes way beyond how features work, but rather how much data gets sent to the server, to the client, how frequently etc.
A lot of the work in networked games is handled client side - meaning stuff that that the server doesn't care about or needs to know about. Some of that includes things like a player in a doorway, in a ship, shooting, jumping, waving etc. All the play cares about is what a client is doing and where they are in the game world (2D or 3D).
e.g. in Star Citizen, when you have 8 players in ArcCorp, each one needs to know where they are, and what they are doing. So if playerA is waving, PlayerB needs to see that. In a peer-to-peer game, where all clients talk to each other, that data is communicated between clients. In a client-server game, where the server is referee who doesn't trust the client - since they can cheat - it's the server that is responsible for receiving, analyzing, and sending that data to other clients. So if the networking isn't optimized and a lot of data gets sent to the server, which then has to process and send it back, it can cause some significant performance issues.
If you know about Big Benny, then you know the problem. A player can move it, but the other player won't know about it because the server doesn't handle anything that isn't related to actual player clients. And if they do "grabby hands" cargo, it's the same problem because while they would have to process that data specifically, there can still be issues whereby the cargo is lost, dropped, not in the right location etc. All the problems currently seen in 3.0.
All the buzz words you see CIG and backers uttering, like network bind culling, serialized variables etc, aren't even close to what they need to solve their networking problems. Yes, they will help, but their impact is minimal - at best - in the general scheme of things because there are LOT of other
INSURMOUNTABLE things that they have to overcome - and which they can't.
Things like players being in one ship, trying to board another etc, is trivial to do because it's just positional data. It's no different from how you have a player flying around in space, and another driving on the moon below. As long as they are all in the same instanced world, those kinds of things aren't hard to do. But they can be plagued with problems as we have seen in the game since 2.0 was released. And in 3.0, due to the addition of new things and features, it just got worse. And there is just no recovery because there isn't enough time to rip it all out and do it right. They waited too long to do that - and they're hamstrung with their use of the WRONG engine for the game they are trying to build. Using on variant (Lumberyard) of CE3, isn't going to solve it for them. Even AMZ has stated that the engine isn't for MMO games. Sure, you can use it to build one, but you would have to write a LOT of support code for it.
All that aside, trying to do inter-instance connectivity (like how ED does it), is going to be worse because of how the game works. ED gets away with it because they built a robust networking engine from the ground up, then plugged it into their game engine. They planned for it from the start. And they only have ships to deal with. Star Citizen on the other hand has a LOT of features to deal with in a networked world, including the fact networking an FPS game is fraught with issues all by itself.
I was on an Open House stream yesterday in which I mentioned that backers have not realized or reconciled the fact that, in the 400 people they claim to have working on this game, NONE of them has EVER worked on a game of this scope, let alone an MMO. So they've basically been doing R&D, as learn-as-you-go. That's just not going to work, and that's why the project has suffered, and is on the verge of a catastrophic collapse.
More here:
http://dereksmart.com/forums/reply/5949/