A sudden interview with an unexpected programmer
At Skies' Edge is a bolt from the blue flight arcade game that seemingly came out of nowhere in May 2023. After months of playing every single update and writing a few articles, I found myself with more questions than answers. Because it has a hard to track its development history, Skyward Flight Media reached out to its developer, Mackerel Sky, who turned out to be a very interesting fish.
Thanks so much for accepting my interview request. Your project has been something of a mystery to me. I have a lot of questions.
Thanks for reaching out for this interview! My name is Mackerel Sky. By day I am a fish, but sometimes at night I turn into something that could be vaguely construed as a programmer-like lifeform. At Skies' Edge is also a mystery to me and hopefully this interview gives me an opportunity to think about what I'm planning for the future.
How did you become a fan of flight games/simulators?
I’ve always loved aviation and flight – as a kid I was always getting my long-suffering mother to draw me pictures of airliners, jet fighters and bombers so I guess gravitating towards flight games was an inevitability.
Eventually I got my hands on a PlayStation 2 with Ace Combat 5 on it – I couldn’t understand a word of English at the time but it clicked eventually and I've been playing this stuff ever since. I stick very closely to the flight arcade end of things, as I’m only here to see cool planes going fast and exploding things.
When did you start considering developing your own games?
If we’re talking being involved in games generally, I guess you would have to start with Vector Thrust. I was in high school when that was being smashed out and did some of the draft writing and many, many aircraft descriptions for that game. There’s also Project Sandwall as well, but we’ll talk about that later on. I wouldn’t really count these as “developing” anything.
I would say I only really started looking at gamedev towards the latter half of my university days, so I guess that would have been about 3 or 4 years ago. Around the time somebody said I wasn't going to be successful if I kept playing or being involved with videogames. I overreacted a little and started looking up tutorials a few days later. My programming experience also has paid dividends at my day job, so I guess in a small way I proved them wrong.
That being said, I’m definitely more of a 2D artist than a programmer, so my first project, designed to teach myself C# was a little pixel-art platformer called Sable Hearts. It eventually grew way too big and taught me some valuable lessons in scope creep. You could say ASE is actually my second project.
At Skies’ Edge has a hard to track development history. From what I could dig up, everything can be traced back to something called Project Sandwall. What was it?
Project Sandwall was the working name for a flight arcade engine thing developed by another student. I signed up to do a lot of the things I did on Vector Thrust. My memory is really bad so I don't recall much - but the gist is that we tried to turn it into an actual game, got hit with the cold reality of game development (using a custom engine, no less) and eventually we both kind of lost interest in the project – other interests, jobs and priorities came up and things ended quietly. For me especially, I learned that having just a vision won't get you anywhere - you really need the skills to back it up.
When Project Sandwall ended, were any of the builds saved? Any assets saved?
I think I have a build of it floating around, but it needed an active server to connect to. I commissioned a few 3D ground assets that I stuck into ASE- for example, the AA gun, APC and airfield models are all leftovers from Project Sandwall.
Around the same time in 2017-2018, Project Wingman was in early active development. Did you ever consider joining their team after Sandwall ended?
There was not much interaction between us off the top of my head. I went straight into the complex and varied kinds of work that fish do after my degree and didn’t really have enough time to dedicate to much of anything for many years, so it never crossed my head.
At Skies’ Edge had a public demo release on May 1st, 2023. That is about a six-year window between projects. What work did you do to get a new project ready? How long did it take?
The gap is about 6 years, but I reckon I spent about 3 years doing nothing, and 3 years slowly learning to program and working on Sable Hearts. You could say that the itch to return to the flight action genre struck maybe February or March 2023, and I took a hiatus on the platformer to play around with some flight sim projects made publicly available. I eventually thought back to my days on Project Sandwall and decided to see if I could modify Lunetis’s OperationZero project to fix some bugs and have it function more like how I would have envisioned the final flight model and HUD to be.
About a month into screwing with it, a buddy in a community I frequent was moaning about the lack of PC flight games and how new projects get announced and die quickly (like Project Sandwall, I guess). I posted some videos, making very clear that I was going to go back to Sable Hearts once I got bored, but was promptly dogpiled and very heavily encouraged to finish what would turn into v0.01a, so I got permission from Lunetis to keep using his codebase, and by May I had something I felt was presentable.
ASE is firmly a flight arcade title, but it has some notable aspects of its flight model that I think make it standout compared to its contemporaries. Is there anything you are proud of with the way aircraft handle in this game?
I wouldn’t say I’m particularly proud of anything flight-model wise, but I did want to try to distinguish ASE from Ace Combat and Project Wingman. Aircraft have a set ACM (Air Combat Maneuvering) speed range where they’re most agile. Instead of complex physics calculations, I use a graph that allows me to directly define maneuvering ability as a function of speed. This allows me to set areas of high and low performance very easily for different aircraft. The manual throttle system is designed to help players find and stay in ACM speeds more easily.
I’m actually most satisfied by the throttle controls, which allow you to instantly jump to full or minimum power by double tapping accelerate or brake. Where aircraft are capable of supermaneuvrability, double tap and holding brake deactivates the limiter, and releasing it restores the regular flight model. I think it’s really intuitive to use in combat, and I’m planning to enhance the visual and audio feedback for these actions further down the line.
Aircraft weapon loadouts are notoriously hard for flight arcade games. There is always the risk of providing too many options and too much ammunition to players, to the point it throws off gameplay. How did you settle on ASE’s method of handling weapon selection?
I will be honest here and say I really haven’t done much work on this apart from getting the basic skeleton of things up and running. Currently, ASE uses a hardpoint based system where aircraft can mount any weapon in their inventory to a hardpoint strong enough to support that weapon’s weight. I have the following weight categories: Light, Medium, Heavy, Super Heavy.
Hardpoints can support multiple categories of weapon weights and also can have specific weapons banned from them if they need to. Multi-lock and multi-launch is dictated by the weapon itself.
Weapon selection was a topic of detailed and passionate debate during the early design stages of ASE v0.01a and v0.02a, and I still haven't decided on how I'm going to handle the final version of it. I will also need to take into consideration how the rest of the game is going to be designed, like how many units are going to be in the map and how weapons perform. This is really only a placeholder system until I figure out what I really want. I’d like a system that allows a selection of multiple weapons but is balanced enough to encourage players to select the right aircraft and loadout for the job.
I appreciate that in the flavor text for aircraft and weapons, there are mentions of countries and manufacturers. Is there a possible world and lore expansion planned?
ASE kind of fell out of the blue and I haven’t done any work on this at all. I briefly considered going back to what I wrote for Project Sandwall, but I took a look and it was pretty disgusting edgy teenager slop, so I pulled out some names to use and then deleted the rest. The names will also probably eventually be replaced, but I did want to show that this was eventually going to have a story.
If there is anything I’ve learned over time, it’s best not to be creative until you know what your foundation is. So I’m holding off on doing any detailed planning until I know exactly how ASE will function mechanically, but I do want to create a campaign, and I have an overall vision of how the game will feel and the background of the lore/world.
The tone of the story will be hopeful. My vague idea is to explore the challenges and conflicts involved in rebuilding a world that was taken to the brink of mutually assured destruction.
This game used Operation Maverick / Operation Zero by Lunetis as a base. Has your recent months of game development pushed your skills farther?
I guess you can say that I'm developing my skills - the vast majority of what I touch or add now is new, which is good for learning, not so good for actually making something work. As things progress, they'll also get more complex, so development may slow down as I work through what I need to do. I'm also conscious that Lunetis has done a lot of the hard work for me, which is great in that I don't have to worry about that stuff but also concerning because I wouldn't know how to reconstruct things like the flight model or mission structure from scratch.
The five missions that are currently available represent what is possible with this game. But these are not something solely produced out of the project from Lunetis. The development blog for update 0.05a discusses new AI management processes and a mission manager system. Why are these two updates so important? Could this lead to larger developments in the long term?
Lunetis’s Operation series were tightly scoped, highly tailored projects designed to replicate specific scenarios as closely as possible given the resources available to them.
In particular, the code behind the mission manager and AI are custom written for those scenarios, where the player is the only allied unit on the map and must destroy some targets. However, ASE is not scoped as small as Operation. This has meant that the AI and mission manager need to be upgraded to be more flexible and respond to changing situations on the battlefield.
Before 0.05a, I had done a bit of work to the AI and mission managers to bash them into something that could go beyond their Operation implementations, but quickly realised that this was a bandage solution that was inefficient and inflexible, so I had to step back and really smash out a framework that could give the AI enough flexibility that it could be placed in any mission and still execute its objectives correctly. Fundamentally the core of the AI hasn't changed, it's just now able to better interact with the mission manager to support a wider range of objectives, allowing me to make much more complex scenarios without having to custom-code anything specific like in Operation. Importantly, the AI and mission manager are modular and expandable, so if I need a new behaviour, I can write it up, integrate it with the framework and don't need to change anything else.
The game has a handful of updates since the demo was released. Some of them have been pretty substantial. When the demo dropped, what was your original plan? Did the reception of ASE change your plan?
After being told to finish v0.01a, my plan was to drop it and then go back to Sable Hearts. I was not prepared for the demo to be picked up and shared amongst a lot of people, which really picked up after Skyward FM found and posted about it. The attention definitely played a factor in convincing me to keep going and people gave me a lot of encouragement.
Around the same time, Sable Hearts was quickly swept away in the never-ending flood of 2D pixel art platformers, and to be honest a platformer was not the best medium to tell the story I wanted to for that game.
These two factors combined to push me to keep going on ASE, and I’ll go as far as I can since I’ve still got that guiding vision in my head. I'll admit, Sable Hearts is the story I really want to tell, and I'm keeping it in the back of my mind for when I'm skilled enough to do it justice, just as how I’ve done for Project Sandwall.
…Ouroboros?!
Uh-oh. Sounds like the indie dev cycle! Haha! Today, do you consider yourself a game developer or is this something you do as a pastime?
I am just a fish who has been lucky enough to enjoy a particularly good work life balance, and I have a number of other work, education, and social commitments outside of ASE. Game development is just a hobby to me – something I did half out of spite and half to learn a new skill. I’m not sure if I would do it full time, and it’s part of the reason why I won’t work on the art or story until I know what ASE is capable of. I don’t want to get people’s hopes up or spend any money unless it’s ready for it.
I appreciate your time in clarifying a lot of this. Personally, I find myself enjoying At Skies’ Edge quite a bit for what it is. Thank you very much for this interview.
Thanks for inviting me to talk about ASE. I've been told to include some behind the scenes screenshots for this interview but they're all quite boring to look at. I've added in some previews of what's coming next in 0.06a to help spice things up.
About the Writer
Aaron "Ribbon-Blue" Mendoza
Co-founder of Skyward Flight Media. After founding Electrosphere.info, the first English Ace Combat database, he has been involved in creating flight game-related websites, communities, and events since 2005. He explores past and present flight games and simulators with his extensive collection of game consoles and computers. Read Staff Profile.
.png)
.png){kind=logo}
