SPOILER-FREE: Data-mining yields good news

[VlonaldKerman]

WARNING: This post contains UNVERIFIED information concerning the development state of KSP 2 and therefore is SPECULATIVE in nature.

 

In the recent past, I've expressed some resigned cynicism and concern about KSP 2, given a number of things. I won't rehash all of that now, and given the information we had at the time, I think my concerns were very reasonable. However, I just came across this reddit thread: https://www.reddit.com/r/KerbalSpaceProgram/comments/11caznb/psa_from_a_dataminer_stop_calling_the_devs_lazy/, in which a person claims to have found a lot of promising things in the code. The spoiler-free, tldr, is this:

Many (most? all?) of the key selling-point features of KSP 2 are in fact present in the code of the current build in an unfinished, but relatively advanced state. It seems as though the intention was to not have an early access, but a full release, at some point in the future, and thus, all of these features (including the ones in the current build) were worked on simultaneously, with the intention of them all being finished around the same time. Presumably, at some point before last October, an edict from T2 came down that something was to be released by March 2023, and so they scrambled to finish certain features more in order to push out a premature EA release. In other words: contrary to many other EA releases, which focus on perfecting more the core gameplay experience before touching the other, more advanced features, in order to release into EA in a timely manner, the intention wasn't, until recently, to release an EA at all, and so a smaller proportion of the devs' time was spent on the stuff that we are playing with now than is typical of EA releases.

There is also another thread: https://www.reddit.com/r/KerbalSpaceProgram/comments/11b8s6f/comment/j9wocuy/?utm_medium=android_app&utm_source=share in which someone claims to have found that the physics engine in KSP 2 has in fact been rewritten, from scratch.

Please correct me if I've made any inaccuracies here- I wrote this up as quickly as I could after coming upon this information. If it's true, I think it's important for this content to be up on the forum, as many people, myself included, prefer this forum over the other places in which KSP 2 is discussed, and this information gives me huge optimism for the game moving forward, when I was just feeling the most pessimistic that I've ever felt about KSP 2's future.

It's unfortunate that it's not really feasible for the devs to come clean about the bureaucratic process that seems to have led to this PR meltdown and botched release, but hopefully we as a community can put the pieces of the puzzle together and spread awareness of the facts concerning how the release came to be, in its current state. That is, assuming these are the facts.

 

Am I wrong? Missing information? Not allowed to discuss things like datamining on this forum? Please let me know before too many people get the wrong ideas! As far as I know, I haven't made any factual errors, but information is spread about in so many places, and I'm currently too busy to do full on journalism. 

GODSPEED KSP 2!!!  

 

 

[RocketRockington]

From that post, it seems like a very cursory inspection of the code.  Changed namespaces doesn't mean much - you can do that in an IDE in no time at all, it's akin to doing find and replace in a word doc.  Similarly, the names for features existing in the code is relatively meaningless without knowing how much code has been written for them, what state it's in, etc.  It's extremely common to have code stubs or assets for features that might be a long time to be delivered, or are even never going to be functional and will get replaced with other things.

Another programmer looked at multiplayer code and found that while there was code to do high level stuff like login, the actual difficult part of making multiplayer work - low level synchronisation code, esp. was missing.  The actual difficult part of making multiplayer work. Network login/handshake you can just buy packages for from the Unity Store and drop them in if you wanted.  I'll try and dig up the reddit link for it.

Again, nothing says here they actually rewrote the physics engine either.  It still seems to be using stock PhysX, with many of the issues present in KSP1, hence unlikely to be in different from KSP1s engine, not in any fundamental way.

I'm not accusing devs of obfuscating development btw.  It's very common when starting a new project using legacy code to re-namespace code that has been reviewed and made up to date even if it's fundamentally the same algorithms, just to find within the new structure of even just coding standard guidelines of the new project.

I just doubt this idea that there's a bunch of unfinished work that is magically going to start working in a brief window of time.   It'll trickle in, but not fast. Primarily because there was a large window of time where the project was slated for delivery within the FY2023 release window.  If they did have to yank and suddenly disable a bunch of stuff at the last second, it's doubtful it was 95% complete just because what they did leave in was also clearly a less than finished product too.

Edited March 2, 2023 by RocketRockington

[HowDoKerbal]

It’s not uncommon for code references for other features to be found in game files, especially for games still in development which KSP2 most definitely is. It’s basically in an alpha state.

Undoubtedly, other features were present for testing and development in whatever dev build we ended up getting and they were simply removed or turned off for the early access release.

[VlonaldKerman]

1 hour ago, HowDoKerbal said:

It’s not uncommon for code references for other features to be found in game files, especially for games still in development which KSP2 most definitely is. It’s basically in an alpha state.

Undoubtedly, other features were present for testing and development in whatever dev build we ended up getting and they were simply removed or turned off for the early access release.

Yes, not surprising this stuff is in there, but encouraging to see it does in fact exist.

[zeekzeek22]

I agree with your positive assesment of what probably happened. It really makes sense. I'm really optimistic that once the core game is like 80% debugged, the rest of the main feature modules will drop in fast succession since a lot of the features are pretty straightforward technically, if you assume everything else is debugged.

[K^2]

18 hours ago, RocketRockington said:

Another programmer looked at multiplayer code and found that while there was code to do high level stuff like login, the actual difficult part of making multiplayer work - low level synchronisation code, esp. was missing.  The actual difficult part of making multiplayer work. Network login/handshake you can just buy packages for from the Unity Store and drop them in if you wanted.  I'll try and dig up the reddit link for it.

This has been discussed long before there was even an anouncement of the early access. We know that Intercept got even their first multiplayer engineer pretty late, and that it's the part that would have been under the highest risk. My expectation was the game shipped without MP at first. This is consistent with it being left in the most unfinishes state. A lot of other systems are much further along. I haven't dug into it too deep yet, but you can see what assemblies are hooked up even by just going through the save files. A good example is how KSC is set up - it's a colony. We might not see any colony editing, but we have a fully functional, if sometimes buggy, implementation of the colony system already shipped. And that's not even the first item on the roadmap.

In short, there are good reasons to expect science, colinies, and interstellar to drop in quick succession, along with critical bug fixes and some QoL fixes and updates, giving us a well rounded experience. The rest will take longer. I suspect it could be closer to half a year or longer before we see multiplayer, but that shouldn't be the yardstick for everything.

[MarcAbaddon]

The devs state that they are already trying out multiplayer though. But it's really not in the code, while science and colonies are to some extent. By the way, I now had a look at the physics and a cursory glance it's mainly wrapper of Unity RigidEngine as I suspected. At least, at some point it lets the RigidBody class (that's the one from Unity) handle stuff.

[RocketRockington]

30 minutes ago, K^2 said:

In short, there are good reasons to expect science, colinies, and interstellar to drop in quick succession, along with critical bug fixes and some QoL fixes and updates, giving us a well rounded experience. The rest will take longer. I suspect it could be closer to half a year or longer before we see multiplayer, but that shouldn't be the yardstick for everything.

I'm less optimistic about the timeline.  I don't think we'll see multiplayer in 2023.  Of course, I'm not basing that on any specific evidence.  I also don't expect we'll see the rest of the roadmap in short order - I expect they'll be solving issues with this EA for at least a couple months, though science (of some sort) should be such low hanging fruit in terms of implementation that if they've got the pieces, that should go in relatively fast.  Colonies and interstellar... maybe by mid-to-late this year.

[MarcAbaddon]

Half a year seems short, yeah. I would be happy if we get one of the items on the roadmap every 3 months, suspect it might well be double that.

[K^2]

1 minute ago, RocketRockington said:

I don't think we'll see multiplayer in 2023.

Unfortunately, plausible. My expectation of us seeing something of it in about half a year is based primarily on how far along everything else is, and them being able to dedicate significant resources to now. But that can be off, resulting in further slips. At the very least, I'm confident they're pushing hard for a 6-8 month timeline, whether it's the most realistic outcome right now or not.

2 minutes ago, RocketRockington said:

I also don't expect we'll see the rest of the roadmap in short order - I expect they'll be solving issues with this EA for at least a couple months, though science (of some sort) should be such low hanging fruit in terms of implementation that if they've got the pieces, that should go in relatively fast.  Colonies and interstellar... maybe by mid-to-late this year.

I've worked on games that looked worse than this early access a month from ship. It looks really bad when you're playing it, but if you actually go over the list of A and B bugs, it's not that bad. The issues they've already mentioned as being addressed for next release cleans up a great deal of game-breaking bugs. Even things like physics, that look really bad, isn't actually in such a terrible state. I've built a number of very stable craft by avoiding certain parts and certain connections. This suggests that most of the Kraken visits are due to a few buggy joints, rather than an entire system being bad. So I'm rather more optimistic overall. But we'll see soon enough.

[RocketRockington]

19 minutes ago, MarcAbaddon said:

Half a year seems short, yeah. I would be happy if we get one of the items on the roadmap every 3 months, suspect it might well be double that.

We're not far off...im.giving them till the end of the year for 3 features after all.  But yeah could be double that.

17 minutes ago, K^2 said:

I've worked on games that looked worse than this early access a month from ship. It looks really bad when you're playing it, but if you actually go over the list of A and B bugs, it's not that bad. The issues they've already mentioned as being addressed for next release cleans up a great deal of game-breaking bugs. Even things like physics, that look really bad, isn't actually in such a terrible state. I've built a number of very stable craft by avoiding certain parts and certain connections. This suggests that most of the Kraken visits are due to a few buggy joints, rather than an entire system being bad. So I'm rather more optimistic overall. But we'll see soon enough.

I've worked on one like that.  But there were a few differences with that title and this one.

1. The game was on a great trajectory.  We saw the build was getting better from week to week.  We ended up delaying 6 months because we knew our release would be great if we did.  (It's current metacritic score.is.sitting at a 90)

2. The bugs were mostly content errors.  Lots of them.  But the core game and toolset was good. It was adapted from the prior version

3. The team was fairly good at planning but the scope was.just somewhat more than we could handle.

 

In some ways this mirrors KSP2's state, and I don't know fully internally what happened, but I'm more pessimistic for them.  Primarily because the bugs are everywhere - both content and key systems.  Because it was shoved out the door early - it's clear noone thought one more month was going to be a magic bullet, which would have still left it technically in the FY03 target.  And because the many year+ delays indicate a team that is struggling or poor at planning.

[Drakenred65]

1 hour ago, K^2 said:

A good example is how KSC is set up - it's a colony. We might not see any colony editing, but we have a fully functional, if sometimes buggy, implementation of the colony system already shipped. And that's not even the first item on the roadmap.

 

And a surprisingly indestructible one at that, given or even with the bugs and glitch's . That said I'm wondering if KSP is planed to upgrade over time since it's a colony only more in the background or as you progress in non sandbox mode.

[K^2]

37 minutes ago, RocketRockington said:

In some ways this mirrors KSP2's state, and I don't know fully internally what happened, but I'm more pessimistic for them.  Primarily because the bugs are everywhere - both content and key systems.  Because it was shoved out the door early - it's clear noone thought one more month was going to be a magic bullet, which would have still left it technically in the FY03 target.  And because the many year+ delays indicate a team that is struggling or poor at planning.

Intercept started working on this game in 2020, and while we got a lot of mixed messaging about how it got delayed over and over, realistically, late '22 was always the target. I'm sure it was internally, and the rest is basically just damage control over the Star Theory situation targeted primarily at the investors.

With that view in mind, we're looking at a delay of just a few months. If the science, colonies, and interstellar land in the next 3 months, and most of the bugs are fixed, then barring the multiplayer component, the rest of the game is on the similar 6 month delay overall. And yes, I think it's primarily due to scope being more than Intercept was anticipating.

There are a few bugs that I'm surprised have made it this far, and suggests some sort of either a rush or poor QA on Intercept's part. Or maybe, if they are relying on Private Division's QA, some communication issues between the two. But outside of that, the core actually seems good. The only two core systems I would highlight as needing an overhaul is the symmetry/part sets system, as it's far more brittle than it should be, and the world origin relocation system - which looks like someone was trying to experiment with what they'll need for multiplayer, and left it unfinished, causing the craft to "drag" other craft and even colonies with them while traveling. Between these two, it covers the majority of mission-ending bugs I've seen people post. Including most of the Kraken visits, as these are tied to symmetry modes being a bit broken when you start moving parts manually or placing struts. The rest consists of a few bad states, some bugs in physics vs logical connection, and a ton of content errors. This is all very fixable.

And yes, I absolutely might be too optimistic about it, because maybe I'm suffering from confirmation bias, but so far, everything is lining up pretty close to the trajectory I was expecting from the time Intercept started publishing their first updates on progress.

P.S. Based on Linkedin profiles, we know that the engineer responsible for the physics changed late in '22. It looks, again, purely from work profiles, that the previous physics engineer left for a better offer, rather than indicating any sort of an internal problem, but someone new coming in and having to learn the setup could be another explanation for why some of the bugs slipped into the core systems so late. The person who took over doesn't have a lot of gamedev experience, but they have a solid background, so I hope they have been ramping up quickly. Obviously, a source of uncertainty, but could also be a more benign explanation to some of the issues we've seen in the early access.

Edited March 2, 2023 by K^2

[RocketRockington]

10 minutes ago, K^2 said:

And yes, I absolutely might be too optimistic about it, because maybe I'm suffering from confirmation bias, but so far, everything is lining up pretty close to the trajectory I was expecting from the time Intercept started publishing their first updates on progress.

I'm glad you can admit that.   But to me, a project that has to announce multiple large delays - even if some of them are damage control, plus is dramatically feature incomplete, missing features that were part of the target from the outset, plus is in a very buggy state at launch that was clearly not well planned... that's a project that has been operating dysfunctionally.  And sure, this moment might be the moment they solve thier dysfunction(s) but that's hardly likely.  Sometimes release pressure does help sort things out - but we'd see that reflected in the launch, I'm sure they've been under pressure since the last slip.    I guess we won't know until we see the ongoing trajectory, but so far the trajectory does not inspire confidence.

[K^2]

5 minutes ago, RocketRockington said:

But to me, a project that has to announce multiple large delays - even if some of them are damage control, plus is dramatically feature incomplete, missing features that were part of the target from the outset, plus is in a very buggy state at launch that was clearly not well planned... that's a project that has been operating dysfunctionally.

My entire point is that this wasn't a dysfunction in the studio. PD committed to a release schedule with Star Theory that got then abandoned along with the Star Theory. Intercept was created as a new studio and allowed to build the game mostly on their own timeline. The fact that PD continued to promise that the game will be released in late '20, then some time in '21, indicates a problem in the publishing and marketing arm of this whole operation, not in the development studio.

If we look at some early updates from Intercept, along with hires they've made early in '20, it's pretty clear that the game's production was started almost from scratch in the spring of 2020. What we see in the early access was delivered in under three years of work by a modestly sized team. This indicates good operation and direction to me. So I don't think they have to fix any internal problems at the Intercept. Just continue doing what they've been doing without distractions or burning the team out.

[GigFiz]

Yeah, I've been a bit puzzled by statements that seem to say how bad it is, because this is the state things are in, which is fair, and that we have practically none of the promised content, which doesn't necessarily say anything about how done the rest of the game is, just that it hasn't been released to us yet; it's enough to be cautious, certainly, and maybe even a bit pessimistic, but the "end is nigh"/all is doomed" is kinda over the top or, at the very least, premature. Stuff we've seen over time, such as the dev diaries, show and tell segments, and whatnot pretty clearly indicated, usually with photo or video evidence, that all the systems that they were telling us about were actively being worked on, potentially at solidly progressed stages (not possible to make a definitive statement here. There was enough evidence to solidly make that argument, but also little enough of it that the counterargument of "maybe that's all they had, and were just cherry picking to make it look further along" is, at the very least, still a valid possibility, which should not be dismissed.

I  do think it is likely that a lot of the game is further along than many people fear, and they went the route of starting with just releasing the core piece, and getting it stable and performant first, then adding another and doing the same (which also gives more time to work on the unreleased stuff in the meantime). I imagine that if a larger percentage was released at once, at this stage of stability, having that many more pieces at once would exponentially increase the complexity of the the job and increase the time between updates, so they opted for stabilizing a piece, then adding and repeating, rather than trying to do it all at once.  Sort of like (for a sloppy metaphor) taking a knotted tangled rope and starting at one end, doing a piece at a time, then moving to the next, rather than just dumping the whole mess in front of us and trying to untangle all of it at once.

 I am absolutely not saying the complaints and issues are not valid, or trying to dismiss the many, real concerns, but also I suspect that early access is going to progress more steadily, and faster, than many people think (I'm not going to guess, just...faster than expected, especially with how pessimistic estimates have gotten. Call it a reasonable and steady pace, not an interminable morass), especially once they get the core pieces sorted out. Now I may be totally wrong, but for what it's worth, the datamining thing here is at least another piece of supporting evidence (not proof. Just another little piece of evidence). Time will tell; the pacing and scope of the first few EA patches and updates will be very telling about how things are, and will go, I think.

[regex]

4 minutes ago, K^2 said:

If we look at some early updates from Intercept, along with hires they've made early in '20, it's pretty clear that the game's production was started almost from scratch in the spring of 2020. What we see in the early access was delivered in under three years of work by a modestly sized team. This indicates good operation and direction to me. So I don't think they have to fix any internal problems at the Intercept. Just continue doing what they've been doing without distractions or burning the team out.

Man, hearing that from an industry veteran inspires a lot of confidence. Understand that I've been a professional developer for nigh on two decades now but business software is a pretty different game, so while I had my own suspicions about the progress here (buggy but otherwise incredibly promising from a team that seems quite competent) it's good stuff to hear.

[Bej Kerman]

12 minutes ago, K^2 said:

My entire point is that this wasn't a dysfunction in the studio. PD committed to a release schedule with Star Theory that got then abandoned along with the Star Theory. Intercept was created as a new studio and allowed to build the game mostly on their own timeline. The fact that PD continued to promise that the game will be released in late '20, then some time in '21, indicates a problem in the publishing and marketing arm of this whole operation, not in the development studio.

If we look at some early updates from Intercept, along with hires they've made early in '20, it's pretty clear that the game's production was started almost from scratch in the spring of 2020. What we see in the early access was delivered in under three years of work by a modestly sized team. This indicates good operation and direction to me. So I don't think they have to fix any internal problems at the Intercept. Just continue doing what they've been doing without distractions or burning the team out.

4 minutes ago, regex said:

Man, hearing that from an industry veteran inspires a lot of confidence. Understand that I've been a professional developer for nigh on two decades now but business software is a pretty different game, so while I had my own suspicions about the progress here (buggy but otherwise incredibly promising from a team that seems quite competent) it's good stuff to hear.

And this is what I've been trying to tell @Alexoff et al. KSP 2 is already big for how young it is, that's not even considering the roadmap features that the devs have been working on over the same three years as the core experience. Glad to hear it from professionals

[RocketRockington]

17 minutes ago, K^2 said:

My entire point is that this wasn't a dysfunction in the studio. PD committed to a release schedule with Star Theory that got then abandoned along with the Star Theory. Intercept was created as a new studio and allowed to build the game mostly on their own timeline. The fact that PD continued to promise that the game will be released in late '20, then some time in '21, indicates a problem in the publishing and marketing arm of this whole operation, not in the development studio.

A couple of counterpoints.  It is a ' new studio' with the same project leadership as the failed studio

You claim this is a PD issue but PD is the one who brought them in and structured this studio.  And if PD is promising bogus dates , then either there is terrible communication between the studio they just founded, of PD is terrible and incompetent but is still running the show there.  Also they had Nate Simpson - an Intercept dev - announcing these dates, not a PD marketting person.

You claim it was a fresh restart - but clearly they took a lot of code and assets from the failed projects, based on the 2019 promo material and the post star theory releases. The VAB looks identical.  The lighting looks identical.  The UI looks identical till they did an internal redo.

You say it was a fresh restart but the entire design team came over, so the project should not be credited with having to do a new concept and pre-production phase, much of that work should either have been already done - or if it was done badly, they pulled over an failed design team.

Overall I think your assessment is the most pollyanna version of events and I don't see any reason to credit them for such a pollyanna version.  I think you're stuck in this fallacy of devs as angels and the publisher as devils.

Edited March 2, 2023 by RocketRockington

[Bej Kerman]

2 minutes ago, RocketRockington said:

I think you're stuck in this fallacy of devs as angels and the publisher as devils.

<s> No, because everyone is a devil apparently  </s>

[RocketRockington]

6 minutes ago, Bej Kerman said:

<s> No, because everyone is a devil apparently  </s>

No they're all just people.  And people can screw up.  And when they screw up repeatedly, believing they will stop screwing up in the future is a nice thought but not the most likely outcome.

And they also will follow their incentives - including the incentive to lie to the community and paint the prettiest picture they can about how things are going.

Edited March 2, 2023 by RocketRockington

[Bej Kerman]

Just now, RocketRockington said:

3 minutes ago, Bej Kerman said:

<s> No, because everyone is a devil apparently  </s>

No they're all just people.  And people can love up.  And when they love up repeatedly, believing they will stop loveing up in the future is a nice thought but not the most likely outcome.

For what KSP 2 is after only ~3 years, and given all the background content being worked on at the same time, I wouldn't consider it much of a love-up

[Moons]

A lot of conclusions with almost no information - just because there is code and some file-names doesnt mean nothing ...

Since this game has been developed by another studio before chances are there is also a lot of old code etc. - long story short - i wouldnt draw any conclusions from this.

 

But some people also want to draw conclusions from not released builds etc. - and most of the time people will draw the conclusions they want to have. (not even sure how to look at that or how its named since i dont care ...)

Edited March 2, 2023 by Moons

[Bej Kerman]

1 minute ago, RocketRockington said:

lie to the community

<s> Like the time they lied about the game running perfectly on mid-tier hardware? </s>

[RocketRockington]

7 minutes ago, Bej Kerman said:

<s> Like the time they lied about the game running perfectly on mid-tier hardware? </s>

When they lied saying they needed extra time to make that the case, and either did nothing to make it so in the interim, or weren't in the state where performance was the top problem.

Why are you giving that sarcasm tags, that's exactly the case.

Edited March 2, 2023 by RocketRockington