• Register
Post news RSS An Alternative UI?

The history behind Cogmind's 4:3 UI, and considering whether and how to provide a separate layout for smaller screens.

Posted by on

Normally when I publish a new article, the focus is on explaining the results of some new feature/mechanic/design. Today's post is a little different because it's about an important feature that has not yet been implemented, nor have I come to a decision on how best to handle it.

I'm using this post as a way to lay out all the facts and considerations before making a decision, while at the same time giving anyone else who'd like to provide input an opportunity to do so.

So what precipitated this post?

For some players the Cogmind UI is too dense given the size of their screen. I prefer playing in a small 720p window myself, but when making a game for wider consumption it should be as accessible as possible for as many different types of players as possible, and some players find that their setup results in objects/words that are too small to parse/read clearly--they "just want bigger fonts."

Unfortunately it's not as simple as that.

Cogmind already comes with many fonts, including very large ones, but there is an upper limit to what can be used in combination with a given screen simply because the current UI requires a minimum of 60 rows to display all its information.

Note that I used the term "screen" here, and not "resolution." The latter is not an issue, since Cogmind's font system will allow it to scale up to any resolution with pixel-perfect crispness (right now I've only included full-size font support up to 2160p/4K, but more can easily be added on request).

Thus the question we're addressing here is one of physical screen size--regardless of resolution or pixel density, displaying 60 rows of information means dividing that physical screen height by 60. The only approach, aside from recommending play on a desktop or large-screened laptop, is to offer a different UI layout that reduces the amount of information available at any given time, and apply this across the entirety of the interface. Oh my.

History behind Cogmind's UI

How did we get here in the first place? The current UI's history goes back surprisingly far, to even before Cogmind's conception...

In 2011, I began working on X@COM, a tactical squad game based on the original UFO Defense. Naturally one of the first things that needs to be decided with a game represented via an emulated terminal is the dimensions of the grid. I had spent many years developing games in real terminals before that point, so it felt natural to go with the standard terminal width of 80 characters. As for height, the largest UFO Defense maps play out in a 60x60 grid, so it looked pretty perfect to fit that into the standard width and leave a 20x60 area for the UI sidebar.


A shot of the early X@COM UI (it was later changed, and there are other windows). (Click for full size.)

The following year I decided to participate in 7DRLC2012, and I certainly wasn't going to start from scratch, so I basically gutted the X@COM source to use as a foundation in creating Cogmind 7DRL. The latter ended up being designed around an 80x60 terminal, purely because that's what X@COM was using at the time.

cogmind_7DRLCogmind 7DRL from 2012, which has multiple font options as well. Coming from X@COM I did reduce the map view to 50x50, since maps are much larger than that, anyway, and we needed the space for a log plus more info in the sidebar. (Click for full size.)

So if you weren't already aware, now you know that Cogmind originated as a game jam game that itself splintered off a UFO Defense fan game. The early decisions definitely had an influence on how the game developed, because at no point in all this did I start over from scratch.

When I decided to pick up Cogmind again in 2013 to turn it into a full game, I made a short-lived attempt to start over before deciding it was a waste to throw away all that prior work (this was discussed on the blog a couple years ago).

In terms of UI layout, for Cogmind the 60-row minimum was just enough to fit all the information I wanted to include in the 7DRL, and font size was also less of an issue then because all characters were square, and therefore could use twice as much horizontal space as in the new version, making them easier to read even when small.

I was really excited to transition from the 7DRL to the modern version capable of narrow fonts, effectively giving me a 160x60 terminal where non-map text is concerned, because it meant I could fit in all the extra information I needed to make the UI as useful as possible with little window switching. (I hate square fonts, anyway, so I was happy to see them go.)

And there you have it, the origin of Cogmind's default 4:3 aspect ratio, which is somewhat out of place in today's world of HD resolutions. Of course I've made the UI dynamically expandable, allowing the logs to fill as much space as possible, though this can often result in a lot of black space up top since log messages were meant to be short and the contents of those windows were designed to fit well within their narrower default width. (See the end of this post for diagrams/explanation regarding UI expansion.)

Thus we finally come to the current version:

cogmind_UI_1920x1080_tiles_cogmindCogmind's current primary UI, as it appears in the most commonly used resolution, 1920x1080 (click for full size). Currently the second most commonly used resolution is 1366x768.

I've also put together a new reference page where you can preview what Cogmind looks like at different resolutions using different fonts, especially useful for those prospective players who aren't sure whether Cogmind's UI will be too dense for their particular screen of choice.

Why an Alternative?

Now that we have our background and recognize the current state of affairs, the first important question to look at is why it's even necessary to explore an alternative. There are two main parts to this...

The first is a question of marketability.

These days, Let's Plays on YouTube and Twitch make up one of the most vital marketing channels for indie games. (It's a fairly obvious trend; see this report for one summary with data to back it up.) If a game isn't easy to stream/record, or doesn't look so good in that environment, it's less likely to be picked up by as many channels. A UI crammed with fine details doesn't hold up against compression very well, and won't always look great unless viewers are willing and able to watch in a large window at high quality. When I stream, I somewhat cheat on this by using my 1200p screen to stream at 1080p, letting the lower 120 pixels be cut off since that area's not too important for viewers, anyway (and therefore I can stream using a larger font).

So a very dense UI is non-ideal for recording or streaming, and while Cogmind has already received fairly good video coverage since launch (from smaller channels!), reception and exposure would no doubt increase with an option to enlarge the UI elements.

Interestingly, I did recently encounter an unexpected marketing-related benefit from Cogmind's UI density when it appeared in a 2015 GotY list alongside other games like Undertale, Helldivers, and Everybody's Gone to the Rapture, where the editors couldn't do the game justice with a smaller image like they could with the other games, so they just gave Cogmind a much bigger image :P.

The second, more important question is one of individual accessibility.

As mentioned before there are plenty of players on laptops rather than desktops (the second most frequently used resolution is 1366x768, a common netbook resolution) as well as older players and others with poor eyesight that need to be considered.

And again, it's not so much a question of resolution, but the fact that physical screen sizes are shrinking compared to years past (even compared to when Cogmind was first developed).

I've been contacted about this literally dozens of times since Alpha 1 released. The number is only a small portion of players overall, though there is a common rule that for every person who makes a statement, unless it's something completely off the wall then you can expect there are X more thinking the same thing who just didn't speak up about it. That and alpha players are more forgiving. Add these X's up and it's no longer such a tiny portion of players.

My first response was to release alternative fonts, which was enough to satisfy some of these players. However, that only works for those for whom the default text was perhaps just a tad too small or, more accurately, was not as readable due to the stylization in combination with its small size on some screens.

We need to examine the feasibility of a more heavy-handed approach.

Roadblocks & Solutions

If only the solution were easy! But no, there are some major roadblocks in the way here.

Anyone who's followed the blog for a while, or played Cogmind before, will know that a lot of attention has been given to the UI design, and it's unfortunate that the ideal solution here--reduce the amount of information available at any given time, would undermine years of careful thought and effort.

Showing less information at once breaks one of the fundamental rules in the design, that as much vital information and interactive elements as possible be accessible at once, without additional windows or scrolling. The UI has been quite successful* at this, though it ends up requiring a large grid to achieve it, leading to the topic of this post. (*One notable exception is the meta strategy some alpha players are using with large inventories, which require excessive scrolling and will likely prompt me to add a new inventory management mode to alleviate that down the road and fall in line with the usability of the rest of the UI.)

So at the root of the problem, having started out with 80x60 terminal means that each UI element's size, relative location, contents, interactions, everything even including game mechanics have been influenced by that decision. It's a delicate house of cards in which any piece that's moved or resized will trigger a ripple effect that essentially lowers the quality of the entire game (unless it's redesigned from the botto--NOPE :P). If I'd taken this issue into consideration from the start, Cogmind would be a different experience!

However, here I must admit that as part of my vision for the game, I was intentionally cramming the interface with as much data as possible. It does a very good job of fulfilling the goal of creating a sometimes overwhelmingly detailed environment that really gives you the feeling that "You are the Cogmind" :D. In that way it's been a success, just not a success that transfers well to all devices...

Let's begin investigating specifics!

(Images for discussion below will use Cogmind's "native" 720p mode, all of which can be opened to see them in their crisper full-size form.)

First of all, note that because Cogmind was designed to a 4:3 aspect ratio, our limiting factor is height--width is never an issue, what we need to do is reduce the number of rows of data in the interface. How many rows we can afford to lose, and what impact does that have on the rest of the UI?

Starting with the main interface, let's assume we're redesigning it 1) for a 16:9 layout and 2) to allow for an increase in relative font size across all resolutions. My first thought was to move the inventory from where it's hanging out at the bottom, since that is the biggest potentially movable chunk...

cogmind_ui_concept_move_invConcept: Moving the inventory to the top.

Doing so would give us back 14 rows! Sure enough, such a change would allow the current 1366x768 players to upgrade from a size 12 font to size 16, and 1920x1080p players to move from size 18 to size 22. However, such a radical change leads to multiple problems on the main UI alone...

  • It would reduce the amount of visible inventory from 10 items to 6 in order to fit the top area.
  • It drops the viewable map height from 50 to 36, meaning that with the player at the center only 18 spaces are visible to the north/south, while it's technically possible to see out to 24 spaces according to the mechanics (not to mention all the other mechanics that work beyond 18 cells and are normally within view).
  • The current inventory position relative to the parts list above it is actually quite important. Conceptually it's a smaller area below (= less important than) the list of attached parts, and to use something you move it upwards (= raise its importance). (Yes, this was part of the original UI concept, and these are the kinds of things I think about when working with a GUI--many of my ideas come from prior training in linguistics and conceptual metaphors.) Moving the inventory away also loses its parallelism with the parts, both of which share the same info visualization modes to allow for more direct comparison of values like integrity or other qualities.
  • A lot of the game is spent moving items between the parts and inventory windows, and further separating them would slow that process, especially for mouse players.
  • If the inventory were to become a part of the top-center multi-console, as it's the most commonly used window the player wouldn't be as free to choose one of the other optional windows as desired (the area was intended for situational secondary windows).

So, um, yeah, that's not gonna work.

Before we continue experimenting with the main UI, let's look at the effect of changing the console height on other windows. We do have the status window, item info, hacking interface, and the help and many game menu pages to think about. These, too, were naturally all designed with the 60-row limit in mind, along with the desire to have as much information available as possible without scrolling or forcing separate windows.

cogmind_ui_status_dataSample status and data windows, for reference.

Without removing any existing data, these two windows have about 3-4 lines to spare. Ideally we'd also want to leave an empty row at the bottom because it looks better. Given these examples I can see ways to condense the amount of data to free up a total of 5-6 lines without any serious sacrifices in content or quality. Technically there are a few items that manage to fill their entire data window, so those would need to be dealt with as well, but they're in the minority. (No scrolling allowed here!)

The hacking interface has plenty of room and flexibility, so no need to examine that.

As for game menu pages, most of them are dynamic and don't required their full height--basic commands, manual, game, news. The options and credits pages each have about 8-10 rows to spare at the bottom. The art gallery would need a slight redesign to be able to free up at least 8-10 rows, but that one's not too difficult. The only one that would need a bit more work is the advanced commands page, which is already packed in its current form :/

cogmind_ui_advanced_commandsThe advanced commands help, for reference.

Looking at the numbers overall, the scale of the 14-row inventory move described earlier would also be an impossibility where many other windows are concerned. Something lesser might be feasible, though.

Before we continue, here it's worth bringing up a common misconception among players just starting out: It feels like there's plenty of empty space in the HUD, so why not just allow the contents to be larger? Or, for example, why not shorten the parts list by the necessary amount since that area's not being used?

Anyone who's played longer knows that, as you can see in the sample complete UI image shown earlier, that list eventually gets filled up with parts--26 of them from A to Z. There's actually not a single row to spare, and seeing that list in its entirety is extremely important, i.e. not something that can be sacrificed to scrollability. That number of parts was another mechanic that was decided early on but could've easily gone another way had I decided it was necessary to have a lower part limit for the purpose of shrinking the UI. Now the entire game is balanced around it!

So let's take another look at the main UI for different candidates from which we can realistically remove/relocate some rows. Within the inventory, there are the two item-name scroll buttons at the top/bottom, empty if there is no previous/next item. We could theoretically get rid of those and enable scrolling via another means.


Concept: Slightly condensed inventory.

That doesn't really work, though, since it would ruin the consistency of the aesthetics across most of the other scrolling windows, most of which are laid out and interacted with in the same manner (plus it looks bad). More to the point, only removing two rows isn't going to make a meaningful enough difference to allow for font size increases.

A better candidate for adjustment is the HUD area above the Parts window, specifically the Scan and Volley windows, which don't have to be at that particular location. In fact, both could be redesigned and provided as info over the map or other windows.

The Scan window's info is actually almost entirely available through other means as is (it's a legacy 7DRL feature left as redundant info even after adding the dynamic map info system we have now). The Volley info is only really used in firing mode, and could be represented on the map, or together with weapons when preparing to fire--it only needs to be shown in that mode.


Concept: Volley data strip (containing hypothetical data just for demonstration).

Unfortunately the 4 rows this saves aren't quite enough--it turns out we need to reclaim at least 6 rows in order to give our most common resolutions the next font size up (1366x768 would increase from 12 to 14, and 1920x1080 from 18 to 20). So we'll still need two rows from elsewhere...

The top HUD area has two empty lines, though that's a bit of intentional black space to keep things from getting overly cramped. Even more important, the HUD is shaped as it is because the corner intersection of the top-side windows and sidebar forms a makeshift rectangular area that doesn't need a border of its own. Shifting the sidebar upward looks bad.


Concept: Shifting the HUD and condensing the upper block.

Well, now that I draw it out, it doesn't look quite so bad. Yay for mockups. I prefer the bi-directional diagonally mirrored look, though the zig-zag intersection is somewhat interesting. The corner layout's density is a bit over-the-top, and it looks kind crushed and less important considering it's actually your most vital info...

But we did free up 6 rows :P.

There would be quite a few other consequences of this and changes mentioned previously, in terms of little UI feedback details and interactivity, but I've been focusing on the big picture here. To point out some of the other problems on the main UI alone, the cascade of changes would also require: moving the volley visualization toggle, a new home for the hidden Scan buttons, some new way of providing the names of machines, another method of showing the item data summary... All of these elements fit and work nicely where they are now (while also making sense!); I think the overall quality of the UI suffers by shuffling these things around, since they've already been optimized, after all.


In terms of difficulty, providing an alternative UI doesn't seem quite as bad as I expected, but at the same time the benefits are not as great--it only enables one step up in font size.

Converting the base terminal area from 80x60 to 80x54 does bring it almost half-way from the 4:3 (1.25:1) aspect ratio we started with to a native 16:9 (1.77:1) better compatible with today's screens.

That said, under the new system non-16:9 screens (the most common one being 1280x1024, approximately 5.7% of Steam users) would still not benefit from the changes, remaining at the same font size. On the technical side, the advantage of a 1.25:1 ratio is that you know you can always expand it horizontally to fit whatever screen space is available, while it's not so straightforward with a terminal with an odd 1.48:1 ratio =p.

On the plus side, even a one-step font increase will slightly narrow the top-side windows back towards their intended/designed width, which has always been stretched too far for these 16:9 screens.

Mapwise, shaving 3 cells (6/2) off the sight range to the north and south is okay, though it would mean we'd probably need to remove the best tier of sight range-enhancing items, simply so they're not annoying by frequently spotting hostiles outside your map view.
(Note: As is the case now, players can always opt to use a smaller font to see more of the map at once.)

Timewise, implementation could require devoting the majority of an entire major version release to this one feature. That's about 3-4 weeks.

Financially I believe the effort could probably still pay for itself down the road by expanding the number of potential players and somewhat mitigating the number of complaints. (I'd be far more certain of this if we could increase the font step by two instead of one, but that appears out of the question.) That said, money has always been the lesser of my concerns, with the priority instead on making an awesome game. Cogmind was, after all, designed from the ground up with these particular UI dimensions, layout, and content in mind, which has in turn heavily impacted the mechanics.

Technically, one of the options here, and how I originally approached the issue, would be to make the alternative UI an optional "special mode," i.e. not the default appearance. But maintaining two such distinct modes is complicated, and would continue to be a drag on future development. Fortunately the 6-row proposal reached above is not a significant divergence from the original UI, thus there's a chance it could become the UI layout.

In summary, it's doable, but do we want to do it?


If the necessary UI adjustments were any larger I was originally just going to chalk it up to lack of foresight and say it'll have to wait for Cogmind 2 where the design could account for it, but based on the analysis above (carried out as I drafted this post :P) it seems worthy of serious consideration.

The question, now more specific, becomes whether it's worth investing the necessary time/money while also making slight sacrifices in the original UI.

One aspect that keeps me from enthusiastically jumping on this opportunity is that the second most common resolution, 1366x768, would be upgraded from Cogmind's "native" (and best-looking) size 12 font to size 14, which is the absolute worst in terms of style (it's an intermediate font size that's difficult to balance properly given the style).

There's also the loss of the half-width HUD windows that nicely divided up the sidebar to keep it from looking too monotonous, not to mention providing effortless access to more information... :(

I've prepared some mockups showing what Cogmind's UI would look like under the new system in the two most common resolutions.

cogmind_ui_concept_alternate_HUD_1366x768_tiles_cogmindConcept: Condensed 1366x768 UI. (Click for full size.) Also in ASCII form, and using one of the alternative fonts (Terminus). Compare to what 1366x768 looks like now.

cogmind_ui_concept_alternate_HUD_1920x1080_tiles_cogmindConcept: Condensed 1920x1080 UI. (Click for full size.) Also in ASCII form, and using one of the alternative fonts (Terminus). Compare to what 1920x1080 looks like now.

What do you think? Is it better? (Worse?) Is it worth it? Is it enough?

Post a comment
Sign in or join with:

Only registered members can share their thoughts. So come on! Join the community today (totally free - or sign in with your social account on the right) and join in the conversation.