Stellaris Dev Diary #154 - UX Design in Stellaris

[Olterin]

Oh yeah, one thing regarding the UI that I've been meaning to post: could we please either have sorting options on the resettlement window, or automatically put unemployed pops on top of the list?

 

[Anhell]

Hi there, i´m a big fan of Stellaris (ST mod ), and a web UX Designer, and i´ve been always interested in UX/UI game interfaces, so this post was a real nice reading, would it be possible to write in Medium or anywhere you want, an article about these experience, sharing images and videos ??.
Thanks for the snippest so far , it was a grear idea

 

[MichaelJanuary]

Enough has been said about the planet screen, the outliner, notifications, and the need for planet summary screens with editable build orders or decision toggles.

What I would like to talk about are the two things which attracted me to Stellaris in the first place. The galaxy map and the star system map. I would dearly love for more of the game to play out on these two screens rather than through popups or notifications or outlines.

GALAXY MAP
- when mouseover a star (or holding down a key like alt) show the total system resource income including from colonised planets, so the full value of the star system is clear.
- show some empire management info like crime levels, stability, happiness, population ... rather than as dry notifications or popups. This will make your empire look like a place with actual life inhabiting it.
- if a system needs attention, or has a notification, show iit with a blinking icon on the map in addition to the notification bar or outliner. I want to feel I am managing a star empire, not a spreadsheet. This can even be a blinking or color coded circle marking the star, and if you mouseover, or zoom closer it shows more information.
- When you mouseover a star system, have an expanded tooltip that lists selectable assets like fleets and planets, so you can click into exactly what you want without having to drill down into the system view and scroll and pan and zoom till you find the right thing.

STAR SYSTEM VIEW
- instead of right clicking objects, can we have a flyout menu as you mouseover things, so you can click commands directly ... e.g terraform, colonise, survey, build station, build megastrycture.
- For colonized planets, have an information overview that pops up without having to open the planet screen (population, crime, stability, hab, current construction if any, with progress bar or time to completion.
- Allow for simple commands to be given without having to open the planet screen, like enabling a policy or decision.

Some of the above information can be zoom sensitive, so as you zoom in or out it filters to show more or less information linked to each star, much like it does with anomalies and fleets.

Basically, let's play the game on the MAP.

Thank you.

 

[Scheffel]

Basically, let's play the game on the MAP.

Thank you.

a Paradox game without at least 15 different windows??? HERESY I TELL YOU!

for a more serious answer...
this system looks nice and all until you have 50 or more alerts at the same time blinking in your screen
because ya know some absolute madman just thought it was a good idea to play the game with "jagji's's big ass map pack"
or even
the in system view with flyout menus would look nice and all until you have some 70K+ fleetpower hovering around that system or the contingency thinks its a good idea to invade your capital as you are enabling things on and off...
i think the current UI works just fine and is easy enough to work with no major problems
tweaks for the view screens are indeed and much welcomed but a whole new UI on top the current one is kinda unecessary.

 

[FlyingPhoenix]

Why on earth would you require an algorithm that gives the optimal result? You can make some very quick heuristic algorithms that will deliver a result from the "good enough" category. In the very simple cases where a human can immediately see how the optimal distribution would look like it would propably give the optimal solution anyway.

I'm not talking about it giving an optimal solution, it just needs to find a near optimal solution and then stop computing.

I'm going to assume that the folks on cs.stackexchange know more than either of us, and here's the answer to the question of whether there exist very quick algorithms which deliver results from the good enough category.:

The TLDR being "No, there isn't really a (significantly) more efficient method" than the Hungarian algorithm.

 

[The Bored Chairman]

Sorry folks, posting error meant the first one got binned. I assure you this is the correct DD in the only canon, official timeline.

That's exactly what a Time Traveler WOULD say! I'm onto you...

 

[Spartakus]

The TLDR being "No, there isn't really a (significantly) more efficient method" than the Hungarian algorithm.

You have noticed that this very answer gives a simple heuristic approach with linear time in it's last paragraph, have you?

 

[MichaelJanuary]

a Paradox game without at least 15 different windows??? HERESY I TELL YOU!

for a more serious answer...
this system looks nice and all until you have 50 or more alerts at the same time blinking in your screen
because ya know some absolute madman just thought it was a good idea to play the game with "jagji's's big ass map pack"
or even
the in system view with flyout menus would look nice and all until you have some 70K+ fleetpower hovering around that system or the contingency thinks its a good idea to invade your capital as you are enabling things on and off...
i think the current UI works just fine and is easy enough to work with no major problems
tweaks for the view screens are indeed and much welcomed but a whole new UI on top the current one is kinda unecessary.

I wouldnt intend for everything to be on all the time. Of course Information overflow is a thing. However, there are such things as map modes, overlays, context sensitive menus etc. You wouldnt list all the ships, but you could have a list of fleet names as a flyout from the ships icon, so you can easily select one fleet and give it a move order.

Not saying I have all the answers, or that everything can be done, but you can take a quick glance at a few other games and see how much more info there is immediately at hand to the player. In Civ for e.g., you dont need a city outliner, you can tell at glance from the map which cities are idle, or what they are producing, or what the pop levels and happiness levels are, or how fast its growing.

Some examples. Currently each star shows the resource income from stations only. (Call it a resource overlay).

If you could switch that to a population overlay, it would instead show (rulers, specialists, workers, slaves, unemployed) for example. And you might have a species filter if you only wanted to see the distribution of one species.

If you had an 'Economy' overlay, it would show total jobs, open jobs, dom servants, unemployed pops ... instead of resource income.

 

[Kain2K]

Big announcement at PDXcon? Huh, can't wait for them to announce massive new features/reworks/DLCs/etc that will break the game again and make it worse while nothing is being fixed. Again.
Fix your game damnit. Or at least don't radio silence us and acknowledge the massive problems going on and give us some intel about the fact you're working on it. Because from where we stand, it seems you simply don't care at all. Adress those issues. Modders are doing a better job than you on this at the moment.
Fix. Your. Game.

Yeah, this really is stupid. While I am curious what that "big announcement" is, I much rather would like to have a solid game, than another expansion, that adds new systems to the game without interlocking with the other DLCs, and without fixing any of the balancing and bug issues we have.

 

[sillyrobot]

You have noticed that this very answer gives a simple heuristic approach with linear time in it's last paragraph, have you?

It's not linear time. n is linear, but only because you add t as a secondary factor. Assuming t stays less than n (not a good assumption for small n) then you approach n squared.

 

[Spartakus]

It's not linear time. n is linear, but only because you add t as a secondary factor. Assuming t stays less than n (not a good assumption for small n) then you approach n squared.

Yeah, luckily we don't care about small n at all. And since t can be freely choosend to offset computation time vs. quality, and we already know our starting graph is much better than a random one - thanks to the whole thing being an iterative process - we can live with a pretty small t.

What really hurts the performance is if I understand correctly is that pops tend to look for better jobs on a daily basis. It would be perfectly fine to queue up all your planets and only reassign jobs on one planet per day. The offset is that if you have 100 planets it might take a little over 3 months to get the job assignment done on a selected world, but if I have 100 planets I tend to not care about them any more indidually.

 

[Tanya Degurechaff]

Hello all, this is Hildor Anduv, the UX Designer on Stellaris. I’ve been with the team since December, and have really enjoyed working on such a great game. With my first Dev Diary I’m going to be sharing how I have been approaching UX Design on Stellaris.

Disclaimer: We're not quite ready to announce our plans just yet (we're saving that for the big stage, PDXCon), so this is going to be more process oriented without delving into specifics about any new feature. If that's not your thing, you can safely give it a miss, but if you're interested in the nuts and bolts of how games are designed and made, let us lift the curtain a bit.

Whiteboarding: The UX Corner
Whiteboards are a great place for getting your ideas out quickly. When whiteboarding I try to get team members from multiple disciplines and rough out ideas, together. It’s a great way to discover gameplay and player needs, as well as find any potential development pitfalls early on.

We wanted to create a permanent space for this, and thus the UX Corner was born. Here we come together and iterate on ideas in real-time as a group. Having the permanent space has helped facilitate conversion and increased cross-discipline communication within the team,.

Wireframing & Prototyping.
From the whiteboard sketches, I then move onto creating more detailed wireframes, the skeletal mockup of a UI design. Keeping your wireframes simple lets you iterate quickly, but more importantly allows you to nail down what content should be displayed, and how best to lay it out visually for the player.

I typically try to nail down one key screen, and from there I build out a User flow of the entire experience.

When Designing the Archeology UI for Ancient Relics, I had some key pillars I wanted to nail.

• The stories are linked, and I can revisit past chapters for context. (Also experience the great artwork again!)
• When I come back here, I have a sense of the progress my scientist has made.
• The challenge of the site feels ‘game-y’.

Once I nailed an overall layout for the Archaeology UI, I then broke it out into a User Flow to capture the full breadth of a player’s journey while excavating a site.

With the User flow complete I then link it up and create an interactive Prototype. The prototype allows team members to experience what we are trying to create before we set out to develop it. It’s a great low cost way to identify issues in the flow, and fix them here before investing in development.

After we agree on the UX design, we then get to finally create it in game. As you can see below, the design mostly held through to release. The text window was increased to accommodate larger story text, and ‘difficulty’ was changed to ‘breakthrough chance’ to help better understand the ‘die roll mechanic’.

User Testing
So, even though the design has now gone through many gates, we must test our designs with players and iterate based on player feedback. Luckily Paradox has a great User Research department, and so we were able to run playtests and get vital feedback which we could respond to before launch.

----------

Well, I hope this was an informative window into how UX Design is done here on Stellaris.

Also, I read the forums and look out for your feedback as well. You all have been very helpful, so thank you for that!

Hello to you, developers! Listen, your game has a graphical problem, which greatly interferes with playing. And here's the problem: when a previously saved game is launched, the “glow” around planets with a unique texture disappears. The unique texture includes two classes of worlds: Earth and Ecumenopolis. Friends, please tell me - can you fix this defect soon, please? It's about the game - Stellaris. Please respond on behalf of the developers.

 

[Tanya Degurechaff]

Hello all, this is Hildor Anduv, the UX Designer on Stellaris. I’ve been with the team since December, and have really enjoyed working on such a great game. With my first Dev Diary I’m going to be sharing how I have been approaching UX Design on Stellaris.

Disclaimer: We're not quite ready to announce our plans just yet (we're saving that for the big stage, PDXCon), so this is going to be more process oriented without delving into specifics about any new feature. If that's not your thing, you can safely give it a miss, but if you're interested in the nuts and bolts of how games are designed and made, let us lift the curtain a bit.

Whiteboarding: The UX Corner
Whiteboards are a great place for getting your ideas out quickly. When whiteboarding I try to get team members from multiple disciplines and rough out ideas, together. It’s a great way to discover gameplay and player needs, as well as find any potential development pitfalls early on.

We wanted to create a permanent space for this, and thus the UX Corner was born. Here we come together and iterate on ideas in real-time as a group. Having the permanent space has helped facilitate conversion and increased cross-discipline communication within the team,.

Wireframing & Prototyping.
From the whiteboard sketches, I then move onto creating more detailed wireframes, the skeletal mockup of a UI design. Keeping your wireframes simple lets you iterate quickly, but more importantly allows you to nail down what content should be displayed, and how best to lay it out visually for the player.

I typically try to nail down one key screen, and from there I build out a User flow of the entire experience.

When Designing the Archeology UI for Ancient Relics, I had some key pillars I wanted to nail.

• The stories are linked, and I can revisit past chapters for context. (Also experience the great artwork again!)
• When I come back here, I have a sense of the progress my scientist has made.
• The challenge of the site feels ‘game-y’.

Once I nailed an overall layout for the Archaeology UI, I then broke it out into a User Flow to capture the full breadth of a player’s journey while excavating a site.

With the User flow complete I then link it up and create an interactive Prototype. The prototype allows team members to experience what we are trying to create before we set out to develop it. It’s a great low cost way to identify issues in the flow, and fix them here before investing in development.

After we agree on the UX design, we then get to finally create it in game. As you can see below, the design mostly held through to release. The text window was increased to accommodate larger story text, and ‘difficulty’ was changed to ‘breakthrough chance’ to help better understand the ‘die roll mechanic’.

User Testing
So, even though the design has now gone through many gates, we must test our designs with players and iterate based on player feedback. Luckily Paradox has a great User Research department, and so we were able to run playtests and get vital feedback which we could respond to before launch.

----------

Well, I hope this was an informative window into how UX Design is done here on Stellaris.

Also, I read the forums and look out for your feedback as well. You all have been very helpful, so thank you for that!

Hello to you, developers! Listen, your game has a graphical problem, which greatly interferes with playing. And here's the problem: when a previously saved game is launched, the “glow” around planets with a unique texture disappears. The unique texture includes two classes of worlds: Earth and Ecumenopolis. Friends, please tell me - can you fix this defect soon, please? It's about the game - Stellaris. Please respond on behalf of the developers. (this english text,dont delet! :С)

 

[sillyrobot]

Yeah, luckily we don't care about small n at all. And since t can be freely choosend to offset computation time vs. quality, and we already know our starting graph is much better than a random one - thanks to the whole thing being an iterative process - we can live with a pretty small t.

What really hurts the performance is if I understand correctly is that pops tend to look for better jobs on a daily basis. It would be perfectly fine to queue up all your planets and only reassign jobs on one planet per day. The offset is that if you have 100 planets it might take a little over 3 months to get the job assignment done on a selected world, but if I have 100 planets I tend to not care about them any more indidually.

There are a lot of possible strategies to alleviate some portion of the current bottlenecks. But claiming there is a O(n) solution when it is patently not doesn't help.

From a design perspective, I'd prefer if jobs remained in stasis unless and until something changed wrt pop count, pop composition, or job count or composition and then were recalculated exactly once. A typical growing colony with a total growth factor of 16 would get two recalculations per year from pop growth and others as dictated by empire construction. I'd also prefer to make assignments from a colony point of view as opposed to a pop -- it's easier to make a few educated guesses that would generally improve initial state / reduce iteration count for a solution.

What I'd do doesn't matter, however. What matters is the experience improves.

 

[Janx14]

Yeah, this really is stupid. While I am curious what that "big announcement" is, I much rather would like to have a solid game, than another expansion, that adds new systems to the game without interlocking with the other DLCs, and without fixing any of the balancing and bug issues we have.

I regret buying the last DLC as I honestly haven't played it at all given that the game's still in the same sorry state its been since 2.2 (which is just adding onto the issues since 2.0).

I don't see why I have any reason to be excited for a big announcement of content when the content we have doesnt work properly. They could announce evereything I've ever wanted for stellaris at pdxcon, but my response would simply be "yeah but how many years till its playable."

 

[Methone]

Yeah, this really is stupid. While I am curious what that "big announcement" is, I much rather would like to have a solid game, than another expansion, that adds new systems to the game without interlocking with the other DLCs, and without fixing any of the balancing and bug issues we have.

Indeed. I'd be entirely happy with the next DLC being another species pack - we're overdue for one anyway, cough cough, aquatics - and while there's an argument to be made about how 'You don't hang up new furniture while taking out support beams', saying you should finish Diplomacy and THEN bugfix, there's also the argument of 'you don't take out support beams while the house is on fire'.

 

[Spartakus]

There are a lot of possible strategies to alleviate some portion of the current bottlenecks. But claiming there is a O(n) solution when it is patently not doesn't help.

From a design perspective, I'd prefer if jobs remained in stasis unless and until something changed wrt pop count, pop composition, or job count or composition and then were recalculated exactly once. A typical growing colony with a total growth factor of 16 would get two recalculations per year from pop growth and others as dictated by empire construction. I'd also prefer to make assignments from a colony point of view as opposed to a pop -- it's easier to make a few educated guesses that would generally improve initial state / reduce iteration count for a solution.

What I'd do doesn't matter, however. What matters is the experience improves.

Apparently we have very different views about what kind of algorithms patently exist or not exist. I completely agree with you however that job assignment should be a lot more stable then it currently is. Unfortunately the AI seems to rely on job shuffling to get through a resource crisis IIRC. But that seems to be an insufficient approach to me anyway.

 

[Yaskaleh]

Great Insight.

Thank you for this enlightening dev diary into the inner workings of how you at Paradox work. Keep up the good work .

 

[Tech Noir Synth]

I would simply pay money for a new portrait pack and biological ships if you just finally do something about the horrible balance since 2.3. We have been posting the same stuff over and over for months. Buff Hiveminds and spiritualists like you did it with Machines in 2.2.5. Rework Hivemind civics. If you do this, everyone will be able to enjoy a game full of diversity and options again, instead of having Machine empires and Synth Ascencion dominate every single game since 2.2.6/ but especially 2.3.

 

[Xenophopos]

I regret buying the last DLC as I honestly haven't played it at all given that the game's still in the same sorry state its been since 2.2 (which is just adding onto the issues since 2.0).

I don't see why I have any reason to be excited for a big announcement of content when the content we have doesnt work properly. They could announce evereything I've ever wanted for stellaris at pdxcon, but my response would simply be "yeah but how many years till its playable."

Indeed, i'm not going to buy anything until they fix the game. Also becouse when you buy a DLC doesn't mean that the supposed DLC will be like they presented you in the first place or will work either. For example, at the begining the DLC "Distant Stars" gave you much more content, more RR and etc. Now you'll be lucky to found the ninites in the L-Cluster (for 5/6 runs i didn't found any trace of nanites and i even have a save to confirm for atleast 1time, but i don't think would be difficoult to have others)... So again, until they finally decide to fix the game (buildings moving around, payedfeaturs missing etc) i won't buy any other DLC, and sure, i'm not exited by another big revelation(=big game breaking). I'm terrified by it.