Star Citizen Live: All Things UI

First they'd ask: why do people want to customise things? Their first thought is that it's because the current UI isn't working for people, so then they'd need to know what people don't like. In relation to this, they're reworking the HUD, Multi-Function Displays, and the FPS visor. Jared posits that one of the reasons people want to be able to change the HUD colour is because it can be difficult to read at times, so being able to change the colour could help with that, but also sometimes people just like different colours, or it could even be related to colour blindness. He continues by saying that it's important to get the basic building blocks of UI done first, before they implement customisation. Zane talks about how they're moving away from having static UI building blocks, to much more flexible ones that should help in solving UI issues. Regarding changing colours to make things easier to read, Zane suggests that there might be a solution to that problem that goes beyond UI (this gets talked about later). Jared reiterates that whilst they're still in development, they want to make the best default standard UI possible, which requires everyone to be using the same thing so that the feedback is unified rather than skewered, because otherwise the feedback would be in regards to specific customisations that would be hard to follow (because there'd be so many different setups). Simon specifies that they are interested in getting feedback about problems people are having with UI right now, and encourages people to share that feedback with them.

Zane says that when they have easily flexible layouts, they can start thinking more about how to make things smarter regarding when info is displayed. He brings up an example that CR has referred to previously this year: the "Target Box" which often just sits there saying "No Target", saying that if you don't have a target, why don't they just not have the box there at all. Typically that extends across all UIs too, such as the MFDs, where if you don't need to view something at that time, it can disappear out of view and reappear when they're needed, i.e. being more contextual. Simon adds that as they've been developing, they've been adding more and more things, so now they can take a step back and figure out what they want to show and how best to do that. They refer to the Chat box as an example of how it's almost constantly overlapping other UI elements when really it needs its own space on the screen, which Zane says is because the Chat box and the Ship UIs are on two different contexts, so they don't know about each other's sizes. He mentions how they're reworking the MFDs right now, creating a whole new system that's much more systemic, where it has a grid system "where everything kind of fits into each other and you can create different sizes of widgets", and suggests that that's probably how it'd be when they revamp the rest of the UI too. They want to be able to not have information repeated, such as being on the helmet visor as well as being on an MFD, as well as Players then being able to specify what information/HUD element they want on their visor. This requires building the foundation first into the overall design, which is something they're spending a lot of time working on right now as the tech is being developed in parallel. So they're building the tools to help them as developers to make really good UI, and then when that's done they can put the good UI into the game for everyone to see and use it, and hopefully to give feedback on it. Jared hints that the next quarter of ISC will show off some of this UI work that's in-progress.

Simon thinks that it's a legacy feature from having so many things added to the game; that they're all competing for attention. It's something they're planning on investigating - the visor display generally - which is: what do we do to this messaging system to make it work for people. For example, they could give messages priority, so that something really important could override other stuff. They could implement timers so that they stay on for a set time. They could restrict how many messages show within a certain amount of time. Zane adds that they need a smarter system that knows what messages should be displayed there, and suggests that maybe the missions could just be a pulsating icon with a number on it, so that Players can see how many unread new missions there are at a glance. Simon adds that they want to split the mission notifications so that messages are in one place, and that they have a mission objective area showing the current mission related stuff, and somewhere separate perhaps for keyboard shortcut hints that might pop up, essentially trying to avoid having them compete for the same space on the screen.

They used Flash and are still using it now, but they're trying to transition away from it by baking their assets into a much more data-driven system. Previously in Flash, you set things up in a static way where you can't see how it looks in-game until you export it and reload the editor. With the data-driven system, the interface with the game code is much more simplified, and they have a standard API so that a task such as creating UI for an ammo counter is really simple and updates live. This means that they can be in-game and have an editor open and work on the UI at the same time, to see what it looks like as they work on it. It's still using Scaleform as a renderer, meaning that they've cut out the process of authoring the UI in Flash, but they're still using Scaleform to draw the vectors but only for that rendering task, and the rest of the work is done by their own code instead. At some point though, they'll build their own render to replace Scaleform. Jared jokingly asks Zane how long he's been waiting to kill Flash, who says he's been waiting since he started working at the company, 6+ years ago (for those who many not know, Zane was an early hire straight out of college and originally worked in the Austin office, which back then was just a house, and this was also during the days of Wingman's Hangar). Simon adds to the discussion by explaining that Flash was originally designed as an animation system for web stuff, so it can be used for things like UI but it's hard to do iterative work with it where things change based upon feedback, which is because it's time extensive. Conversely, with the building blocks stuff it's quick and simple and they have a lot of control. Zane goes on to say that that's also true because it's a fully data-driven system, so the UI is programmatically drawn and driven from data so they don't worry about artists stepping on each others toes, because the changes the artists make can be merged together.

According to Simon, this is another situation where you need to understand the reason for wanting it, and that there may be other ways of solving the problem without creating a compass widget. He suggests that a personal radar or mini-map could show the Player their direction. However, he says that as they revise the visor HUD (which they're starting at the moment) if it seems that a compass is necessary then they'd consider and investigate it. Zane adds that it's not out of the question, but they would need a way to define what's North/East/South/West on these procedural planets, with Jared suggesting they'd have to be able to create and position magnetic poles. Simon suggests that some sort of Sat-Nav system may also solve the problem. Jared adds though that implementing a compass ribbon would be more than just UI work anyway, as it would need to involve system design.

They get a lot of Developers asking about how to do this. Zane says that as they overhaul their UI, they'll be improving the Director Cam system, so it's something they'll take into account at that time, especially since it'd also be helpful for the Devs. He suggests though that it could go further, such as being able to choose what you want to hide, like only the visor HUD, or maybe to hide all of the UI but not what's in the environment that brings it to life - like "background fluff screens". Simon adds that for the general UI, especially the visor HUD for FPS gameplay, they want to have a system which only shows you the UI that's relevant for you at the time, so if you put your weapon away you wouldn't need to see the weapon UI in the corner of your screen, or maybe you wouldn't show Health unless you get injured. This kind of work, which will result in only showing things when they're needed, should also help to de-clutter the screen.

They were basically developed under the hood as "different contexts", and in their overhaul of the design they're factoring in all of the flight information so that it'll be available to Players regardless of what mode they're in, because it's still relevant when they're still flying, and therefore the information should be retained. In these modes referenced in the question, they're looking at potentially contextually changing out the "screens of cells" so that rather than the HUD changing for the different modes, they can shift elements around so that relevant information can still be displayed, rather than the new HUD for that mode just taking up the whole screen. Jared adds that it wasn't intentional to take away crucial information when using these modes, and that it's just something that's happened over the course of development that needs resolving. Simon adds that sometimes you don't realise it's going to cause issues until you try it, and this is one of those situations. Jared goes on to say that things can be developed in isolation, and then when they're integrated together into their game-dev branch, that's when they can see collisions and thus the creation of bugs.

They're looking at a few in-world solutions. The obvious thing to do is to add a drop-shadow to the UI or just make it black, but that somewhat destroys the aesthetic of it. So what they're looking at, which they started looking at with the Gladius but isn't finished yet, is having a system where it's contextually and dynamically reading the brightness of the environment and adjusting the brightness of the UI in response to that. Additionally, to make sure the HUDs look like they exist in-world, they want them displayed on actual geometry to ground it, and they can leverage that to maybe dynamically tint the geometry that the UI is on, as well as then brightning the UI if needs be (depending on the environmental conditions). This solution is ideal because of being in-world (and thus not hindering immersion) but also because it leverages the in-game elements, making it more convincing. Jared asks for clarification, and Zane specifies that it'd involve tinting the physical glass pane but then also brighting the UI, like if you have your phone on automatic brightness and go outside into the sunlight, it'll auto-adjust to make it more readable (This is a thing?! My phone must be old). He adds that it's also an issue with the eye adaptation feature (where the Player's "eyes" adjust depending on how bright it is), because the UI becomes dim when you're on a planet during the daytime, as compared to being lit by the sun in space. They could potentially also have some sort of effect in the UI rendering tech, such as a blurred frosted glass effect, that could help with readability (particularly for the visor HUD in your helmet), and the same is true for busy backgrounds and not just bright ones. Regarding a potential back-shadow, or a black highlight around the words and numbers on the UI (as often suggested by backers) it's definitely something they can consider but it'll depend on how subtle it can or can't be to work, because that might not fit with the aesthetic they're aiming for, which would therefore require them looking for a different solution. Simon adds that as with a lot of the UI, they'll concept different ideas to figure out which is the best way to solve the problem, before committing to implement something.

Yes, that's got to persist at some point, and the issue right now is just that it doesn't. In their new UI tech, the UI will be what they call "stateless", meaning it won't store any state about itself and instead takes everything from an entity or from what's on the server. As such, when they develop the MFD or implement the new design they're working on, they hope they'll be able to persist the current state of the MFDs as they were before the Player exited the seat/cockpit, even if the Player had changed tabs or moved things around. Simon adds that when they do this pass on the MFDs, they want to make sure that the most important information is shown by default, so hopefully there'll be less need to change things around. Zane adds that potentially they'll also be able to make it so that presets could be created, saved, and then activated quickly (he actually just said the "activated" part but that implies creating and saving presets).

Right now the MFDs are small, scaled down, and not readable. Previously they used to have what they called "support screens" which were screens with minimal information on them, with a font size that made the information more legible. They're looking to have a system where by default, the MFDs will be in this minimised configuration where they only show the information that you really need to know, and they do so in a way that's readable without focusing on the display. However, when you then focus on the display, they want it to contextually change to something more in-depth, which can work because now the MFD has more screen-space to show readable information. Zane adds that the cool thing about the UI tech he's helping to develop, is that they're taking cues from web development (which is also his background so he knows a lot about it) where there's a thing called "responsive design". This is where you can have a rule set up so that, if there's a box in their UI that goes beyond a certain point, it then shrinks down, and you can have different styles applied to that, and conditionally so depending on the size of the box. As such they're leveraging that to help with the reformulation of the UI on screens, and it could also be helpful when they potentially implement customisation of HUDs/UIs as a tool to manage and maintain it. Right now in-game they have different sized screens, where each size has its own binary file, meaning that if they want to change one then they have to go into each binary file and make a change. But if they can maintain just one UI, which then has different style rules applied to it, then that makes it much easier to maintain. So changing one thing would then make that change for each different manufacturer, and every kind of configuration. Simon adds that regarding the actual process right now, they're looking through the designs that already exist in-game and working closely with the vehicles team to figure out what they want to show on the screens, to plan out what's going to be in all of the MFDs, so that they can then redesign each screen to achieve its maximum potential based upon what information needs to be shown. After that, the UI tech will eventually reach a point where the screen's are redesigned and the tech's ready to be put into the game.

Zane says that there are situations where the Inner Thought text appears when it shouldn't, such as over screens. A good example of this is when a Player is in an airlock and goes to use the console, and the Inner Thought text then appears over the console despite that same information already existing on the console (it's the same thing as when the door panel reads "Open" but then when you go to use it, the Inner Thought text appears on top of that as well). Their UI tech now allows for not having the IT text appear, which is particularly useful for things like elevators where the required information can be on or next to the buttons, without needing to use Inner Thought. They'll also be looking at the visuals of Inner Thought too, because although it looks okay now they feel it could be better. On a similar note, there's some other work they're doing at the moment on interactions that they're still figuring out, but it's something that's not quite ready to show just yet.

Yes, and it's much more possible now with the UI tech because they have a "style sheet system". Previously (or currently?) this would require having a binary file for each manufacturer, which would be a pain to maintain, but the new UI tech (as mentioned previously) will allow for only one file so that only one change would need to be made to affect everything across the board. Zane explains further that the style sheet system is kind of like having a white box outline which can then have a visual description defined and applied to it, and changing between the different styles is simple because they can just use a drop-down menu to switch between manufacturers, and then see the visual description change between them. Simon adds that once the system is in place it gives them more opportunities to hand it over to the graphic designers who can create really nice designs which would then be a lot easier to just drop into the game, as opposed to being dependent on someone going into Flash and knowing how to code within Flash. Zane adds that with these style rules there are a lot of possibilities to differentiate between manufacturers, but also there are ways to do this through changing the layout, such as Origin and MISC having more holographic UIs. They're also investigating the initial engineering requirement to make it so that they can have 3D UIs as well, which would make holographic UIs look even more holographic. This would be particularly good for the more advanced ships, rather than the more retro ones. Zane comments about how right now every ship just has the retro UI, and that they want to significantly differentiate between the different tech levels of ships.

Yes. The job specs on the website are slightly out of date though and they'll update them soon. They're looking for at least a programmer. They're not currently looking for artists and graphic designers, but that could change in the future. For programmers, the essential things they're looking for are an ability to show experience, and having a knowledge of what makes good UI, such as why things work in other games. For artists they look at a lot of graphic design work because of how relevant that is to UI work, but they also look for an understanding of why a particular screen might be good on a particular app, or how it could be improved. Zane adds that it'd be helpful to have a tools programmer as well. He goes on to say that because the UI is becoming data-driven, that means they're actually dealing with a lot of raw data. As a result, they need to create a UI Editor that the designers and artists will interface with, which would need to be intuitive and easy to use, so a tools programmer that could help with that would be very handy.