Star Citizen Discussion Thread v12

Will they fumble into a working product ? Do we believe in the many monkeys typing on many typewriters theory that there's a non-zero probability they produce the works of Shakespeare ?

Every year that they keep pursuing 'dynamic server meshing', and keep the promised gameplay staples in a holding pattern, it feels like the odds of that get more and more remote. It already feels like the two main paths for the PU are 'janky Second Life sandbox', or a much maligned, but competent, gut and rework.

Currently they’re just running into a wall repeatedly. You don’t accrue much of note that way except rubble.

Who knows on SQ42 though. It’ll probably be a mawkish, old-school single player campaign. They’re not super in fashion. But if they can get it to work 70% of the time I’m sure there’s a market for it. Just probably not a blockbuster one.

Because I mean surely, with all the resources they’ve got, and even if they have borked up the historical mocap, and it’s just a retrograde 'movie experience', surely they can put that together?

Although you look at all the SQ42 badges on the roadmap features...

j6XUQ3K.jpg


And you consider the many years of missed deadlines, and you combine them together to spell 'developmental hell'... And yeah. It does seem possible they’ll balls that up somehow too ;)
 
I'd point out that many other aspects of the core game engine (the client/server main loop, the physics engine, even networking) should have been done in the first few months back in 2012, since that was the whole point of licensing an existing engine. Of course, they could not, as CE is a notorious tangle of vaguely patched together pieces of code that's probably written in German (because why not, SAP are doing that for example) - but again, every small feature they are adding is taking an incredible amount of time, compared to just any other developer around the globe. And then they discard said feature to remake it entirely after understanding their implementation was just not working - and they call that "iterations" (while us in the IT world would call that "complete incompetence"). Will they fumble into a working product ? Do we believe in the many monkeys typing on many typewriters theory that there's a non-zero probability they produce the works of Shakespeare ?
That's spot on and with hindsight, it does point to a very murky business from its inception.
Even with the "cut down" feature set in the kickstarter, any serious analysis of CryEngine produces such a long list of functionality that it just doesn't have out of the box (its many times worse for the current vision), its just not credible to suggest that somehow CiG have stumbled upon the requirements for missing components in the engine whilst they have been developing the game.
For example, CiG claiming they found out they needed to code a 64bit world coord system several years into developing a large scale space game, is like them having 50 pizzas delivered to their offices and trying to claim they didn't know they needed to pay for them with money.
 
For example Unigine 2 have the same features of SE and much more, look and be worried:

ah, it seems that even the new unreal engine 5 will have 64 bit precision.

I've been saying for years that SC should have used an engine like Unigine or any other engine with 64 bit precision out of the box. Even now they should ditch their current franken-engine and change over. Sure it might put them back a year or more (but possibly less)? How many more years will they endure their current entropy-bit precision engine and the countless issues it causes. And management should know not to rely on generalist software engineers to convert it properly to 64-bit precision. Unless they have a team of crack experts in game engine dev, they need to make the hard decision to change over, even this late, while they still have reasonable backer support. At least most of their media assets will survive.

There are compelling reasons games like Dual Universe use Unigine. Even its 2011 release had most of what SC needs, and it supports C++ which SC uses:

"2011-07-21
From now on Unigine can handle as large worlds as required. In addition to effective data streaming the engine supports double precision of coordinates instead of float data type.

It's possible to create virtually unlimited worlds with the highest level of details (maximum coordinates are effectively 536,870,912 times larger). If a single precision world compares to an Earth range, double precision allows to create the whole Solar System and beyond."
 
I'd point out that many other aspects of the core game engine (the client/server main loop, the physics engine, even networking) should have been done in the first few months back in 2012, since that was the whole point of licensing an existing engine. Of course, they could not, as CE is a notorious tangle of vaguely patched together pieces of code that's probably written in German (because why not, SAP are doing that for example) - but again, every small feature they are adding is taking an incredible amount of time, compared to just any other developer around the globe. And then they discard said feature to remake it entirely after understanding their implementation was just not working - and they call that "iterations" (while us in the IT world would call that "complete incompetence"). Will they fumble into a working product ? Do we believe in the many monkeys typing on many typewriters theory that there's a non-zero probability they produce the works of Shakespeare ?

Exactly my thinking too. All of the foundation core engine/software should have been done first. They appear to have been literally fumbling their way through all of the core development for the last 9 years.

It's one of the side effects of crowdfunding, where 99% of the backers don't have an IT background and therefore don't know whether management is successfully producing a game. The second more worrying side effect is that the owners bear absolutely zero financial risk, so faffing around without getting things done actually enhances their personal financial positions.
 
Every year that they keep pursuing 'dynamic server meshing', and keep the promised gameplay staples in a holding pattern, it feels like the odds of that get more and more remote. It already feels like the two main paths for the PU are 'janky Second Life sandbox', or a much maligned, but competent, gut and rework.

On SC server meshing, I'll just point out the Dual Universe home page (using Unigine):

"SINGLE-SERVER TECHNOLOGY
SPACE MMO
No loading, no server instantiation, no tricks. All players share the same persistent universe, at the same time. Dual Universe is the first Metaverse: a common, shared virtual world, controlled by the players."
 
I've been saying for years that SC should have used an engine like Unigine or any other engine with 64 bit precision out of the box. Even now they should ditch their current franken-engine and change over. Sure it might put them back a year or more (but possibly less)? How many more years will they endure their current entropy-bit precision engine and the countless issues it causes. And management should know not to rely on generalist software engineers to convert it properly to 64-bit precision. Unless they have a team of crack experts in game engine dev, they need to make the hard decision to change over, even this late, while they still have reasonable backer support. At least most of their media assets will survive.

There are compelling reasons games like Dual Universe use Unigine. Even its 2011 release had most of what SC needs, and it supports C++ which SC uses:

"2011-07-21
From now on Unigine can handle as large worlds as required. In addition to effective data streaming the engine supports double precision of coordinates instead of float data type.

It's possible to create virtually unlimited worlds with the highest level of details (maximum coordinates are effectively 536,870,912 times larger). If a single precision world compares to an Earth range, double precision allows to create the whole Solar System and beyond."
If they would ve done it back when it first became apparent that Cryengine is illsuited for the task it would be ancient history by now, backers would have accepted and even supported the decision and SC might be on track with a real foundation. Not sure why Chris didnt make that cut, maybe he couldnt due to legal obligations or maybe he simply didnt see it or thought he doesnt need to? I mean in hindsight its very likely that poor old Chris never understood the real scope and challenge of what he proposed.

Can they still do it?

I guess they could even tho it would reset progress back to zero, add at least another 15 years to development and triple the total cost. Tho Chris aint the youngest buck in the stable anymore and I doubt hes cut out for the stress. Needless to say such a move would almost guarantee his place among the biggest jokes in video game history ever ^^
 

Viajero

Volunteer Moderator
That's spot on and with hindsight, it does point to a very murky business from its inception.
Even with the "cut down" feature set in the kickstarter, any serious analysis of CryEngine produces such a long list of functionality that it just doesn't have out of the box (its many times worse for the current vision), its just not credible to suggest that somehow CiG have stumbled upon the requirements for missing components in the engine whilst they have been developing the game.
For example, CiG claiming they found out they needed to code a 64bit world coord system several years into developing a large scale space game, is like them having 50 pizzas delivered to their offices and trying to claim they didn't know they needed to pay for them with money.

On the 64b thing, to be fair, CIG was aware they needed it at least from the early days of Arena Commander and their tiny maps with boundaries, if you remember. 64b was not a consequence of the december 2015 decision to go full planets. They were allegedly working on it at least as early as december 2014:

As we have mentioned before, Star Citizen (and even Squadron 42) presents a challenge in terms of detail and scale that no game has tackled successfully to date. To do what the game requires there needs to be a different approach to how things are organized, rendered and updated. This is why we spent eight months converting the engine to 64 bit precision and why we have developed some new technologies like the Zone system and local grids, which fundamentally change how the engine organizes, streams, updates and renders objects in the world (or more accurately: the ‘verse)

Now, I personally do not think they have succeeded at it. The omnipresent glitches, physics bugs related to position, clipping, the limitation to quantum travel where pilots can only go point to point and no free fly allowed etc etc etc it all suggests a very defficient and quite limited implementation of 64b precision. IANAD but after 6+ years of work on the feature (probably more), and seeing how other games such as Space Engineers, Infinity Battlescape or ED have achieved it with no fuss, in the case of SC it does not look at all like the product of a particularly competent developer.
 
Last edited:
For what was envisioned at the KS Cry Engine was fine. As the project morphed into the expanded vision it was manifestly not the right engine. Changing engine at any point is a bold move, but could CIG have known that the revenue stream would hold up. We can say now looking back, yes it most likely would. Unlike a traditional KS where you have your money up front and can budget your development upfront and consume from it, CIG have an ongoing revenue stream that drives their development budget, so their model is much more dependent on the ongoing revenue stream than would be the case in a normal KS. I believe the engine should have been changed the moment the scope moved, but I can understand why they didn't and the longer its been left the more difficult that decision is to make. I think the decision was also hindered by the fact that CR it would appear prefers to hear its possible from his reports rather than no it's not. This is why I suspect we have all these miracle techs that stem from trying to provide ways to make it work.
 
Every year that they keep pursuing 'dynamic server meshing', and keep the promised gameplay staples in a holding pattern, it feels like the odds of that get more and more remote. It already feels like the two main paths for the PU are 'janky Second Life sandbox', or a much maligned, but competent, gut and rework.

Currently they’re just running into a wall repeatedly. You don’t accrue much of note that way except rubble.

Who knows on SQ42 though. It’ll probably be a mawkish, old-school single player campaign. They’re not super in fashion. But if they can get it to work 70% of the time I’m sure there’s a market for it. Just probably not a blockbuster one.

Because I mean surely, with all the resources they’ve got, and even if they have borked up the historical mocap, and it’s just a retrograde 'movie experience', surely they can put that together?

Although you look at all the SQ42 badges on the roadmap features...

j6XUQ3K.jpg


And you consider the many years of missed deadlines, and you combine them together to spell 'developmental hell'... And yeah. It does seem possible they’ll balls that up somehow too ;)
Every time I see that roadmap and look at all the "gameplay" elements , I wonder when they are going to start on some real gameplay features.

Also, one thought, will there be a way to "level up" in SC? What is the measure/mechanism of progression ?
 
On SC server meshing, I'll just point out the Dual Universe home page (using Unigine):

(...)
Some big differences with DU though: the CEO is actually competent (he's coming from one of the top schools in France), had a solid idea on how to develop the core MMO networking engine (which is the most critical part), used a custom built game engine (see how it doesnt take billion$ to do it), and had a solid scope that was clearly delimited from the start. So they had a chance to successfully deliver, which they did from what i hear, at least in beta.

If they would ve done it back when it first became apparent that Cryengine is illsuited for the task it would be ancient history by now, backers would have accepted and even supported the decision and SC might be on track with a real foundation. Not sure why Chris didnt make that cut, maybe he couldnt due to legal obligations or maybe he simply didnt see it or thought he doesnt need to? I mean in hindsight its very likely that poor old Chris never understood the real scope and challenge of what he proposed.
I see a combination of reasons here:
  • He's really incompetent at software development (which was pointed out more or less directly by Warren Spector & Richard Garriott)
  • As a manager who has taken that decision, he will never back away as it would be perceived as a sign of weakness (management 101, yeah that's what most of them in big companies have been taught)
  • He's poached Crytek employees and tries to make the best of it. That's a backwards reasoning too but one i've seen in many places especially in IT, where managers have a hiring paralysis due to total incompetence on the subject so they'll stick to whatever old crummy piece of software their current staff can maintain, ignoring the pool of actually competent developers currently on the market.
  • He's got Ortwin on board, who knows what deal he's passed with that person

He's indeed bitten way more than he could chew, but I dont see that affecting his decisions much at all (see how he promptly resumed trying to finance his hollywood aspirations).
As for the legal apect, he's got a CE license but it only defines limitations (no redistribution for instance), he could use whatever engine he wants, if he could afford to fire Crytek employees and Ortwin. I doubt he could recruit anyone on the market though as his company reputation is at joke level.

Like poaching/hiring employees of Crytek ?
See what they delivered in 9 years of intense development: nothing. I could list again for you all the core engine aspects that could have been done (probably twice) in that time frame but that would be futile.

On the 64b thing, to be fair, CIG was aware they needed it at least from the early days of Arena Commander and their tiny maps with boundaries, if you remember. 64b was not a consequence of the december 2015 decision to go full planets. They were allegedly working on it at least as early as december 2014:
That's 2 years into the project. Anyone with 2 connected neurons would understand the need of a special coordinate system from day 1. What did they exactly do during these 2 years ? IIRC they sold pixels for $$, and put up a quick CE demo with the base "hello world" example with existing assets and that was the "hangar module". Interesting to see what this thing has become now (it's totally abandoned), or how that hangar would be "connected to the universe" somehow, which will never happen now since they chose to spawn the ships directly from stations or spaceports.

Now, I personally do not think they have succeeded at it. The omnipresent glitches, physics bugs related to position, clipping, the limitation to quantum travel where pilots can only go point to point and no free fly allowed etc etc etc it all suggests a very defficient and quite limited implementation of 64b precision. IANAD but after 6+ years of work on the feature (probably more), and seeing how other games such as Space Engineers, Infinity Battlescape or ED have achieved it with no fuss, in the case of SC it does not look at all like the product of a particularly competent developer.
It's worse than that: from what I can see, they use an euclidean coordinate system for their "interconnected maps", which creates all sorts of issues, that cannot be ignored with such distances. The way to go is of course relativistic frames of reference, which are not only correct in terms of physics, but also solve a lot of coordinate issues, since the transition between two coordinate references would happen in deep space anyway (so you would not notice the associated jerk). Also, polar coordinates. Current system creates a lot of troubles for just the HUD scales to align properly for instance (since where you are on the surface on the planet will change your angle from the current orthogonal coordinates..)
As for QT it's kinda free fly you can stop wherever you want, or free fly anywhere, it works with their "interconnected maps" system. Imagine the star system itself has its own 3D coordinates, notice how all planets are exactly on the same plane by the way...
 
Last edited:
Top Bottom