Stellaris Dev Diary #140 - 2.2.x post-launch patch v2

[Kayden_II]

The development process itself, never snaps end to end. It never goes straight from point A to B to C.

That's not really true: Each "final" version of Stellaris can be considered as such a point: V1.2.5 (base-game ("bg")), V1.4.1 ("bg" + Leviathan-DLC), V1.6.2 ("bg" + Utopia-DLC), V1.9.1 ("bg" + Synthetic-Dawn-DLC), V2.0.5 ("bg" + Apocalypse-DLC), V2.1.3 ("bg" + Distant-Stars-DLC) and now: V2.2.X ("bg" + MegaCorp-DLC). It's actually up to PDS, that each of these "final" versions are 01. completed (in reference to (for example) "ToDo"-code-lines), 02. fixed (in reference to the overall "bug-freeness" and yeah, I know, that this is debatable, but you should get, what I mean) and 03. optimized (in reference to the overall performance, the state of the AI and UI). I could even go a step further: It's actually up to PDS, that each of the (initial) versions of Stellaris are done AT RELEASE: V1.0 ("bg"), V1.3 ("bg" + Leviathan-DLC), V1.5 ("bg" + Utopia-DLC), V1.8 ("bg" + Synthetic-Dawn-DLC), V2.0 ("bg" + Apocalypse-DLC), V2.1 ("bg" + Distant-Stars-DLC) and V2.2 ("bg" + MegaCorp-DLC) ...

So elements of certain updates sometimes start development way sooner than others. An example of that would be Martin working on the planetary management rework for 2.2, as soon as 2.0 and Apocalypse were done. And you all know that in that time we also managed to release Distant Stars.

Fine, but I don't think, that you've done yourself or PDS a favour with such a statement since long story short: You've basically admitted, that, (altough V2.0 ("bg" + Apocalypse-DLC) wasn't the "best" (initial) release, too (my opinion)), PDS had immediately begun to work not only on the next DLC (Distant Stars), but on the current one (MegaCorp) as well ...

Now, as we progress and feel that maintaining multiple branches of the game is costing us efficiency and resources, we try to start moving some of our devs to work on the next update, while designing it in a way, that will allow us to still give the issues you've been reporting deserved attention.

What do / does you / PDS actually mean with "deserved" attention ?

 

[AlphaAsh]

At this point, with all the ongoing discussions, the players' feedback etc. a question is to be asked: is there any good reason NOT to revert to manual sectors?
- drawing sectors seems fit for a grand strategy game,
- gamey tactics (if any) can surely be tweaked with scaling soft penalties
- nobody would ever complain about sector bad shapes
- less coding for Pdox as they don't have to figure some way to avoid creating sectors without holes etc in any possible configuration

So if there is a reason it wpuld be nice to know, at least concerned players (not whiners) would either make helpful comments or no comment at all, i sstead of giving solutions to an already obsolete problem

Because giving the player the ability to draw their own sectors does not fix the fact the AI's planetary development is still severely broken. Thus not helping those who don't want to draw their own sectors and want that automation, and also probably wanting a competent AI opponent along with it.

Two years plus and sectors are still ******. No, in fact they're actually even worse, after all the drama and bull**** that went into getting them to where they were in 2.1. Gawd, I want to forget that ****storm.

edit - Censoring cursing.

 

[Duuk]

We are now in a dedicated post-release support phase for 2.2 and that was the team's main focus. Now, as we progress and feel that maintaining multiple branches of the game is costing us efficiency and resources, we try to start moving some of our devs to work on the next update, while designing it in a way, that will allow us to still give the issues you've been reporting deserved attention.

I guess I needed to read this quoted a few times before it hit me.

Maybe the Stellaris team is too ambitious and needs to be pruned back on its roadmap some. When you can casually discuss "multiple branches" of the game, that could very well be an issue. And while I understand that "art people don't code" is a trope as old as game development, I also understand that Paradox isn't a major development studio with 500 employees so many of the staff is shared and sourced across different games, so if Stellaris isn't using the art team 80 hours a week, then another team will be happy to bill them accordingly. (That's obviously a decision for the suits line Dnote, who we've had the pleasure of interacting with, and he seems like a decent enough suit to work with!)

Maybe the Stellaris team NEEDS to be told to have an "active", "upcoming" and "future planning only - non coding" status teams and to allow no deviation from those status marks, so that active status hours are billed on bug-fixing, maintenance, upkeep, and live-code task work, "upcoming" is working on whatever the current/next expansion or DLC is, and "future planning" is the whiteboard and mockup stage that is being psuedocoded only with no developer time being dedicated. Were I managing the team, I would keep the teams rotating so that once the DLC went live, the devs that were working on the "upcoming" DLC would become the "active" team and the previous "active" team would then rotate to become the new "upcoming" team that would take over the "futuring planning" project, which would obviously become the NEW "upcoming" project. That way nobody is ever "too far ahead", the team always has a focus on current quality, and there is never, ever a belief that "we'll fix that issue later. (Like sectors, which never, ever seem to work right and are always a "fix it tomorrow" problem.)

 

[Serenity84]

is there any good reason NOT to revert to manual sectors?

Depends on what they have planned in the future. Fixed sectors are great with internal politics where sector governors have ambitions and different levels of autonomy. They could cause problems for you to manage and eventually try to secede in civil wars

 

[Balesir]

Sectors are a larger task that is a big too big to tackle right now, with a lot of other fixes being done, but I do want you to know I have sectors as one of my highest priorities going forwards. We will probably not be going back to manual sector management, but we will definitely be looking into how to make the automatic assignment better, while also making sure that resource management and AI development works better. We will probably make a dev diary about our plans for the Sectors in the future, so that we can share our thoughts before committing to anything.

Something hopefully simple to add would be a rerun of the assignment. What I mean is that I have sometimes seen sectors assigned for particularly nice worlds that get colonised early, but the older/core sectors then grow to reach or even surround them (depending what the hyper-net looks like). It would be nice just to trigger a reassessment - done just as if the world was getting newly colonized - for such worlds, incorporating them into sectors they would have been in if they had simply been colonised at a different time.

 

[HighKings_Ruin]

I guess I needed to read this quoted a few times before it hit me.

Maybe the Stellaris team is too ambitious and needs to be pruned back on its roadmap some. When you can casually discuss "multiple branches" of the game, that could very well be an issue. And while I understand that "art people don't code" is a trope as old as game development, I also understand that Paradox isn't a major development studio with 500 employees so many of the staff is shared and sourced across different games, so if Stellaris isn't using the art team 80 hours a week, then another team will be happy to bill them accordingly. (That's obviously a decision for the suits line Dnote, who we've had the pleasure of interacting with, and he seems like a decent enough suit to work with!)

Maybe the Stellaris team NEEDS to be told to have an "active", "upcoming" and "future planning only - non coding" status teams and to allow no deviation from those status marks, so that active status hours are billed on bug-fixing, maintenance, upkeep, and live-code task work, "upcoming" is working on whatever the current/next expansion or DLC is, and "future planning" is the whiteboard and mockup stage that is being psuedocoded only with no developer time being dedicated. Were I managing the team, I would keep the teams rotating so that once the DLC went live, the devs that were working on the "upcoming" DLC would become the "active" team and the previous "active" team would then rotate to become the new "upcoming" team that would take over the "futuring planning" project, which would obviously become the NEW "upcoming" project. That way nobody is ever "too far ahead", the team always has a focus on current quality, and there is never, ever a belief that "we'll fix that issue later. (Like sectors, which never, ever seem to work right and are always a "fix it tomorrow" problem.)

One of the reasons Mass Effect: Andromeda was the disaster it was is because the Bioware branch that was working on it didn't realise how ambitious their original goal was until half way through development, along with barely enough time left.

If the team here are doing something similar then they're in dangerous waters.

 

[Darkspysrival]

Over 10 patches down the road and humans are still called cubs, machine worlds make no sound when being clicked on, ring world segments aren't properly built in accordance with their surrounding segments, and hive worlds are as quiet as mice when you click on them.

But hey, at least we got some UI improvements...

 

[Vilcoyote]

Since this morning, I cannot see all pics on the diary, is anyone having the same pb?

Edit : It seems to be me only sorry for the inconceniance

 

[HighKings_Ruin]

Since this morning, I cannot see all pics on the diary, is anyone having the same pb?
@grekulf

Nope.

 

[permanently_afk]

I guess I needed to read this quoted a few times before it hit me.

Let me pound a bit on that quote. It sounds like something which came directly "from the desk" - it sounds like something I'd tell my PO (product owner) to communicate further "up the ladder" into the direction of customer/management.

Each version (in the broadest sense, the difference can be a single bug fix) increases the amount of "penalty" you pay for maintainance and communication. If you have automated testing, it will also spike the amount of effort those systems need to be kept current. We recently had to spend a bit of time to fix some unit testing after it turned out that *someone* didn't do it properly. If you work on them concurrently, this can eat a large part of your time budget. If things go well, a developer will spend 2-3 days out of 5 implementing stuff. If they don't its one out of five. If things go extremely well, and the team is elite, you'll hit a ceiling at 3 full days with the rest spent on QA, testing, reviewing. If someone forces the team to go beyond that ceiling, you WILL (no ifs, no buts, no excuses) pay for that later on.

As you see, actual time spent "at work" and "implementing" differ sharply. If you maintain multiple versions this will diminish sharply, which is why they - correctly - try to avoid this.

Maybe the Stellaris team is too ambitious and needs to be pruned back on its roadmap some. When you can casually discuss "multiple branches" of the game, that could very well be an issue.

Like all developers, they probably love their work. Ardour is nice, but occasionally needs to to pushed into the present. But yes, they need to scale back with the mechanics. Probably do some design-work (as in software-design, classes, interfaces, etc), too.

And while I understand that "art people don't code" is a trope as old as game development

That isn't a "trope", it is simple fact. Anyone proclaiming something else is either disingenious or stupid. I can't do aesthetics for shit, but I wouldn't let the UX guy in our company anywhere near the database interface. Division of labour exists for a reason.

Maybe the Stellaris team NEEDS to be told to have an "active", "upcoming" and "future planning only - non coding" status teams and to allow no deviation from those status marks, so that active status hours are billed on bug-fixing, maintenance, upkeep, and live-code task work, "upcoming" is working on whatever the current/next expansion or DLC is, and "future planning" is the whiteboard and mockup stage that is being psuedocoded only with no developer time being dedicated. Were I managing the team, I would keep the teams rotating so that once the DLC went live, the devs that were working on the "upcoming" DLC would become the "active" team and the previous "active" team would then rotate to become the new "upcoming" team that would take over the "futuring planning" project, which would obviously become the NEW "upcoming" project. That way nobody is ever "too far ahead", the team always has a focus on current quality, and there is never, ever a belief that "we'll fix that issue later. (Like sectors, which never, ever seem to work right and are always a "fix it tomorrow" problem.)

Eh, that partition probably won't work out - that would require three teams working on one thing. But none of them works on the "entire" thing, they just work on their bit. You'll have a silo-effect before you can say "bugfix release". You want one team, working on the entire thing. That team needs to scale back on "new" mechanics and probably rework stuff we'll never see, but you don't want to split the development team if you have less than nine people.
Also, if you have pseudocode, you might as well do the real thing, since the difference is not that much.

 

[ViolentBeetle]

Dynamically generated sectors, like maybe top 20% planets become capitals and extrude influence could be a solution, but that would also mean sectors disappear and re-appear, who'll take care of the governors?

 

[sillyrobot]

Depends on what they have planned in the future. Fixed sectors are great with internal politics where sector governors have ambitions and different levels of autonomy. They could cause problems for you to manage and eventually try to secede in civil wars

You know when you can alter a function to support a potential future need? In the future when the need is no longer potential.

 

[FiddleSticks96]

Regarding the TA perk: is this really a buff? To me it seems a reduction in the chance to get the techs I would typically be looking for, since most of those are non-rares (like + admin cap, better power plants, bigger stations, bigger ships, megaengineering, anti-grav housing). The only rare that I can remember looking for in the early game, is the psionics...

The tech that reduces alloy cost of upgrading starbases and building starbase modules is pretty stellar. This can save you thousands of alloys, especially if you play wide.

I want to make it clear that this does not mean we will stop making improvements to the game. We will always continue to support our games, but now we need to focus our efforts into a larger patch instead of continuing to deploy smaller patches. The reason why we need to focus on a next, large patch is because trying to maintain multiple branches of development and deploying small patches takes a significant amount of resources away from us working on fixing bugs, improvements and feature development.

So basically, next patch is the end of 2.2.X and going forward we will see 2.3.0, and while bug fixing and performance improvements will still be very much a priority, we will be seeing new "things" in the game, most immediately UI improvements, so you don't waste any manpower? I can be happy with that. I won't pretend to understand how game development works, but efficiency is a concept I can appreciate.

Because giving the player the ability to draw their own sectors does not fix the fact the AI's planetary development is still severely broken. Thus not helping those who don't want to draw their own sectors and want that automation, and also probably wanting a competent AI opponent along with it.

Two years plus and sectors are still ******. No, in fact they're actually even worse, after all the drama and bull**** that went into getting them to where they were in 2.1. Gawd, I want to forget that ****storm.

edit - Censoring cursing.

I'm not sure I can look at sectors objectively anymore. I admit that, between the aforementioned drama and broken sector AI, I have developed an irrational hatred of sectors that borders on a psychosis. They are wholly necessary but I still hate them and still never EVER enable their AI.

Depends on what they have planned in the future. Fixed sectors are great with internal politics where sector governors have ambitions and different levels of autonomy. They could cause problems for you to manage and eventually try to secede in civil wars

That would be pretty neat, but they have to cause good things as well and not just bad things, otherwise, sector governors just become a nuisance. Not sure how this would work with the current ability to "fire" (execute) any non-ruler leader at will. I even make use of this ability when one of my leaders forms an unwanted faction (usually some misguided xenophile faction).

 

[SteelCrow]

A vast improvement! Our tech lead, Moah, has been hard at work improving the way ships are upgraded.

Excellent! If this is what you have in the works, I don't even mind if you boost the cost of retrofitting from how it is now! It was an enlightening experience to have to make real choices over whether or not to upgrade fleets. Hard to say how upgrade time will feel, you might even be able to raise that too without it feeling bad.

Cleaner display of districts! This arcology now display its districts with boxes, and in different colors!

Again excellent! All these new details making things more clear! I am assuming when a planet has more than 9 of Generator/Mining/Farming Districts available, they'll also bring up an expansion box enumerating those districts. Looks like it will be better!

Now, pls return add_building_construction command, or equivalent!

 

[BlackUmbrellas]

So basically, next patch is the end of 2.2.X and going forward we will see 2.3.0, and while bug fixing and performance improvements will still be very much a priority, we will be seeing new "things" in the game, most immediately UI improvements, so you don't waste any manpower? I can be happy with that. I won't pretend to understand how game development works, but efficiency is a concept I can appreciate.

That's the sum of it, yeah, despite all the doomsday predictions floating around.

They're gonna keep fixing stuff, and keep focusing on "fixes first", but they're moving ahead with their existing workflow structure and that means no more tiny micro-patches until the next big patch comes out.

 

[WarStalkeR]

What mod you are using ?

I'd like to know too, that UI is amazing.

My own. Fight For Universe: Infinite Legacy. These screenshots are still work in progress, but will be released in day or two.

 

[HighKings_Ruin]

That's the sum of it, yeah, despite all the doomsday predictions floating around.

They're gonna keep fixing stuff, and keep focusing on "fixes first", but they're moving ahead with their existing workflow structure and that means no more tiny micro-patches until the next big patch comes out.

...This is why I said earlier that we need proper communication, if they're moving on to 2.3 and with a high priority on bug fixes then I didn't get any feel of that from the announcements. The communication doesn't feel clear, if this had been said directly it would have saved a lot of backlash and eased a lot of worries (at least I think it would)

 

[redrum68]

That's the sum of it, yeah, despite all the doomsday predictions floating around.

They're gonna keep fixing stuff, and keep focusing on "fixes first", but they're moving ahead with their existing workflow structure and that means no more tiny micro-patches until the next big patch comes out.

Seems to be the case but kind of have to read between the lines and make some positive assumptions.

My fear even if that is true is that we are kind of stuck with whatever we get in 2.2.5 for a while which at least as it sits now is pretty rough. It will also probably make the beta branch testing/feedback cycle much longer where at least with these small patches we could quickly give feedback on things that were broken.

IMO, the AI and the way it works probably needs close to a rewrite from what I can tell and from what we can mod. Though in the short term they should probably just adopt a lot of the weights/events that Glavius's AI mod does to make at least it playable for folks. Almost all its decisions are very isolated per planet/ship/job/etc instead of cohesive centralized decisions which are then pushed down to its individual elements to execute. They tried to introduce some centralized behavior like budgeting but its not really even that clear how that's suppose to work, if its working properly, and it seems very rigid/static. The AI also really struggles with the market and even doing basic things like buying/selling resources to cover deficits even if it has a large stockpile of one resource and none of another.

 

[kgptzac]

I think it's pretty ballsy to claim 2.2.5 is "good enough" and justify a slower patching cycle. If the devs want some positive hype PR they'd say something like "we've borrowed few folks from other teams to expedite the process of Making Stellaris Great Again", but hey, never admit that you've done poorly at a job seems to be a recurring and despicable theme nowadays.

 

[Badesumofu]

I think it's pretty ballsy to claim 2.2.5 is "good enough" and justify a slower patching cycle. If the devs want some positive hype PR they'd say something like "we've borrowed few folks from other teams to expedite the process of Making Stellaris Great Again", but hey, never admit that you've done poorly at a job seems to be a recurring and despicable theme nowadays.

They haven’t done that, though.