Public Commander Keen Forum

I always assumed that the nonlinear layout in Security Layer was intentional. Gridlock in general does such smart level design that I doubt he left a gaping flaw in this level and just hoped nobody would notice. If you go back to the official games -- especially the Galaxy engine entries -- it's actually rare when a level doesn't have tons of optional branches.

Gridlock in general does such smart level design that I doubt he left a gaping flaw in this level and just hoped nobody would notice.

Lol, I think you're giving me too much credit. I downloaded 1.4, and yeah, it totally is broken right now. The green square switch is supposed to affect both bridges at the same time. You aren't supposed to be able to progress until you hit the switch two floors above, which blocks the way backwards and opens the way forward.

It's been ages since I opened this project (seriously, how has it been over 6 years since I released this!), so I'm not sure when or how this broke. I might be able look into fixing this eventually, though Keen-related projects are unfortunately not a priority for me anymore. I imagine that actually making the fix would probably be very quick, but I'd need to get a project setup going again, which is the more difficult part.

So there's this pretty big indie game called Starbound. Space-themed open-world exploration platformer, seems pretty cool.

In this game one of the primary alien races is called the Avian. They are a species of humanoid birds. From what I've seen their lore is fairly distinct from what is shown of the Krile, but there is a level centered on them called The Great Sovereign Temple which is aesthetically very very similar to The Great Ziggurat in Atroxian Realm.

Gameplay of it on YouTube:
https://youtu.be/fXG4iutQ0_0

Obviously the first similarity to notice is the architecture of ancient golden bricks, followed by the abundance of bird-themed artifacts. The level challenges are similar too. Timed disappearing platforms. A region of the level centered around dodging timed fires, and another area focused on water. Tone of the music is similar as well, especially to some of the HD demo tracks Gamebird uploaded.



Looking at the dates on the forum here, it looks like Atroxian Realm was both publicly announced and then released first, though there was a lot of overlap in development time. Still, I was curious and wanted to ask anyway if there was any inspiration taken from Starbound, or a connection of some kind. (Can't say whether the Starbound devs did or did not take inspiration from AR but I can't exactly just go and ask them on a whim like I can Gridlock hahaha.) Maybe both were derived from the same sources? The similarities are uncanny.

(This game also seems to have some things that very well could have been inspired by the original Commander Keen games, but that's another topic. )

Oh, that's interesting. This is the first I'm hearing of this game, so there's definitely no connection. As to whether the Starbound devs took inspiration from AR... probably not? AR is still a pretty obscure game overall. I've yet to meet anyone else IRL who has heard of it, even among my game dev colleagues and friends (sometimes just Keen is already obscure enough). It would be interesting if they did, though.

AR's biggest influence is Metroid (particularly Metroid II and Super Metroid), which might account for some of the parallels you're seeing. In some ways, AR was an attempt to blend the gameplay of Keen with the atmospheric aesthetic of Metroid. The Krile have pretty big similarities to the Chozo, and that even applies to aspects like the ancient golden brick temples and bird statues. The styles of environments (highly organic, ancient temples, and hi-tech sci fi) also have a lot in common. Even the Atroxians and "Atroxian Realm" could probably be compared to something in Metroid, like the Ing and Dark Aether from Metroid Prime 2. Not that I was trying to rip off Metroid entirely, you can just see how much that franchise inspired me.

When looking at Starbound, I do see the similarity with the music and Great Ziggurat design, though. At least with the Great Ziggurat, that level's design flowed pretty naturally from the idea that I wanted four challenge floors corresponding to keygem colors. And then the obstacles were pretty apparent based on the color (yellow = acid, red = fire, blue = water, green = energy). Although I think the idea of the Great Ziggurat is fairly unique, I've certainly seen plenty of golden brick temples filled with traps in other indie games (particularly Metroidvanias), so it may just be a coincidence.

Ah, I think I read somewhere that Starbound had some influence from Metroid, so that's probably the connection (although now that I'm looking for that info I can't find a source, just ended up finding your post here in the search lol). Never played those games so I wouldn't have noticed.

I know these specific traps are probably pretty common platformer gimmicks, but I was struck by them using the exact same ones in the same kind of level. The fact that it already sort of reminded me of AR and then the similarities kept going further was kinda blowing my mind. I was not expecting another bird alien structure to be a near match for the one that I already knew of, since birds are not well known for building temples in real life, as far as I'm aware.

I guess I was imagining if the Starbound devs had taken inspiration from Atroxian Realm, it would've been something pretty casual, like if they were looking for inspiration from other classic platformers (like perhaps Keen) and maybe stumbled upon some in-dev screenshots of AR and thought "oh that's cool we should do something like that."

Anyway, neat trivia. Thanks for responding. It's cool to think that I can just ask directly if I'm curious about something. Don't usually have that kind of access.

Because there is currently a mesmerizing and delightful playthrough of the game made by TSK, I am vividly reminded of the first time I played AR. I can still clearly picture myself embarking on a long train journey, with my laptop and the game executable packed away as a delightful distraction.

But distraction it was not: From the very first second, the game ensnared me so deeply that the whole world around me disappeared. I remember hearing GameBirds soundtrack for the first time and diving right into those never seen graphics before just like an 8 year old with big bright excited eyes. I immediately got drawn in. AR felt incredibly immersive offering a truly intense gaming experience. The innovations were exceptionally inspiring, igniting a spark of creativity within me.

The thrill and joy of gameplay allowed me to overlook the fact that the level structures were, in truth, rather vast, convoluted. The potential to lose one’s sense of direction—a mistake I often catch myself making as a level designer as well—is perhaps the only flaw in this otherwise magnificent game.

However, aside from that, the distinct visual style that crafts a unique world, along with the myriad clever puzzle sections filled with imaginative riddles, makes this a rich and indispensable entry in the modding scene of the Keen community. You have set new standards and elevated the bar to unprecedented heights. Thank you once again for this incredible product!

Hey yeah I'm also eagerly watching TSK's playthrough (and looking forward to Alphamatic as well ) I did play some AR back in the day but don't think I ever finished it, just watched a bunch of playthroughs because I didn't want to miss out on the story.

I love the overall art direction of this mod. You could argue it's not always very keenish, but neither is, say, Ruin of Roib and I still think it can fairly be called one of the best mods ever. AR has excellent (albeit complex and sometimes purposefully confounding) level design, stellar music and sound, and it really feels like a world you can dive into like Nisaba said. Plus I can't overstate how amazingly well certain details are executed, like the encounters with Atroxus, adding a whole new layer of cinematicness that the original Keen games never had as they were all about gameplay over story. Here you get both! I can't lie, I've never consciously thought about it but looking back I think a lot of my own modding efforts were definitely influenced by AR's ambition. The grim buildup to the final part of the game, the mystery of Directive X_M, the ending reveal, it's all total genius.

Let me go out of my way to praise the music one more time because I'm a freak, I regularly go back to listen to the AR soundtrack on youtube, it's one of the best in any Keen mod, incredibly cohesive and well composed and clearly a labor of love.

Appreciate the kind words! Hard for me to believe that the 10 year anniversary isn't far off. Even though it's been almost as long since I last played it myself, I still go back and listen to the soundtrack often. A remaster of all the music tracks would be such a cool thing to have, but I lack the skill to do it myself.

I wish I had been able to answer some of the lingering mysteries like Directive_X-M, haha. The original sequel concept, Voyage to Veota, is effectively dead. Even though I had a ton of extremely ambitious ideas for it (some of which could probably make for cool standalone mini-mods), it just became too convoluted and impractical for me to figure it all out. I did have some insights recently about a paired down version that could retain some of the core ideas and be feasible to make in an engine like Omnispeak. I'm pretty confident I could make something much better now than when I was in high school/college. But it's pretty tough to justify that kind of commitment, even if I wish I could.

The main thing I'm working on now in my career is a horror game called Fractured Blooms that just got announced at Summer Games Fest. It's been a pretty big opportunity, so that's definitely my primary focus.

Gridlock wrote: Tue Jul 22, 2025 1:14 Appreciate the kind words! Hard for me to believe that the 10 year anniversary isn't far off. Even though it's been almost as long since I last played it myself, I still go back and listen to the soundtrack often. A remaster of all the music tracks would be such a cool thing to have, but I lack the skill to do it myself.

I wish I had been able to answer some of the lingering mysteries like Directive_X-M, haha. The original sequel concept, Voyage to Veota, is effectively dead. Even though I had a ton of extremely ambitious ideas for it (some of which could probably make for cool standalone mini-mods), it just became too convoluted and impractical for me to figure it all out. I did have some insights recently about a paired down version that could retain some of the core ideas and be feasible to make in an engine like Omnispeak. I'm pretty confident I could make something much better now than when I was in high school/college. But it's pretty tough to justify that kind of commitment, even if I wish I could.

The main thing I'm working on now in my career is a horror game called Fractured Blooms that just got announced at Summer Games Fest. It's been a pretty big opportunity, so that's definitely my primary focus.

Personally I cannot speak for Nisaba or K1n9_Duk3, but since she herself has claimed to make it her mission to bring abandoned and unrealized projects to life, maybe you could pass your ideas for the AR sequel to her and see what she could do with it, assuming they aren't too busy with their personal lives and you wouldn't have a problem with someone else creating it. Considering what they did with FitF, I bet they would do an excellent job on VtV. Of course it would be up to you to ask them when you have free time yourself. Why not? The worse that can happen is they say no or they are too busy to commit to such a grand project. It would definitely be something I would like to do myself, but unfortunately I don't possess the skills or the creatively for such a feat, which is why I try not to be too critical of the mods and level packs I don't play or think are crap, simply because I know I could never do better. It's worth a shot to ask, again, when you have the time of course.

I second Syllypryde's comment. Atroxian Realm is an incredible effort, particularly considering it was mostly made by one person, and a proper finale to it has vast potential that would be a terrible thing to throw away. It would likely take a while to make in a different engine as Gridlock initially intended, so a source mod like the original would probably be faster to make. Whatever the case may be, it's provided me with many hours of
enjoyment (at least 50% of which was spent in either Crystalcomb Hive or the Catacombs) so for that, thanks.

The 10th anniversary of AR's release did put me in a Keen mode, so I guess I'll say a few more things that I haven't previously mentioned.

The original plan from 5+ years ago was to build a new Keen engine in Unity with significantly altered gameplay. I envisioned a game with the galaxy art style, but a top-down mode of gameplay with a perspective akin to the early Mario and Luigi games (ex: Superstar Saga on the GBA). It would feature a greater focus on combat over platforming. I had a small prototype, but not much more than that. Alongside that, as mentioned in the Early "Atroxian Realm Sequel Thoughts" thread from a while back, Voyage to Veota was going to feature a chapter-based approach to storytelling that would attempt a more linear game structure.

Ultimately, that vision was both wildly overscoped and kind of an incoherent mess of way too many ideas (even if some were really cool), so I quietly gave up on it. I concluded that a sequel would require a rethink from the ground up, even if some of the core story ideas would still be salvaged.

Maybe once or twice a year since then, I will occasionally revisit the concept and try to come up with a coherent creative vision. It's something I've been unable to figure out for 10+ years. A lot of the challenge comes from the somewhat convoluted connections to the existing Keen storyline and the many, many permutations I've considered. Though, I will say that very recently I may have found out a workable high-level vision, so I guess that's something.

A hypothetical reboot of AR2 would most likely be a modernized Ominspeak-based project, since removing stifling restrictions like memory and tileset limitations would make development much easier. However, getting started has always been an obstacle, as I'm unable to devote the time and energy needed to learn the tech here, and I've been reluctant to ask for help when this project is still mostly a vague pipe dream.

As I've said many times, while it's fun to dream about what could be, the reality of actually making one of the most ambitious Keen projects ever is quite different. This could easily be a 10,000 hour commitment for a niche fan project that maybe 100-200 people will play. That math is hard for me to justify. It would be a lot to ask someone else to take on that commitment, particularly when I'd likely be absent a lot of the time. Though, if there was sufficient interest, I suppose I could keep working on GDD materials and share them eventually.

All that said, I did experiment yesterday with a new title screen for fun. Please don't take this as an announcement that I'm officially making this, but yeah, a new tentative title is "Atroxian Realm: Exegesis".

Gridlock wrote: Wed Mar 30, 2022 2:59

Gridlock in general does such smart level design that I doubt he left a gaping flaw in this level and just hoped nobody would notice.

Lol, I think you're giving me too much credit. I downloaded 1.4, and yeah, it totally is broken right now. The green square switch is supposed to affect both bridges at the same time. You aren't supposed to be able to progress until you hit the switch two floors above, which blocks the way backwards and opens the way forward.

It's been ages since I opened this project (seriously, how has it been over 6 years since I released this!), so I'm not sure when or how this broke. I might be able look into fixing this eventually, though Keen-related projects are unfortunately not a priority for me anymore. I imagine that actually making the fix would probably be very quick, but I'd need to get a project setup going again, which is the more difficult part.

Hey, that's awesome to see that Exegesis might be worked on.

If Atroxian Realm 1.4 is broken, could someone please indicate the last version that is considered 'feature-complete?'

Since 1.2 is what is available on the front page, I will assume that 1.2 is the most recent 'stable release.'

Was that really almost four years ago that I said I was meaning to fix that? How is that possible? Anyway, I checked, and apparently the mistake exists in version 1.2 as well, so you might as well play version 1.4 for now.

I'm really out of the loop with the CK Source configuration that was done for this project, and this was before I got into using source control. I no longer have a working project directory, and I'm actually not sure how to set up an Abiathar project for it since there's no Gamemaps file. If someone does have a working project directory, I could probably make a quick edit and fix the mistake.

In Security Layer, the green square switches are supposed to toggle both bridges simultaneously. That's seemingly not working, but I'd have to inspect the map data to know why. It's likely either a mistake with the infoplane value(s) for the switches, or something to do with continuous bridge tiles between the two bridges (since that's a trick I used to do before it was possible to make a switch target multiple infoplane locations simultaneously).

Gridlock wrote: Thu Feb 05, 2026 4:05 If someone does have a working project directory, I could probably make a quick edit and fix the mistake.

I just compared the Security Layer level in v1.0 to the one in 1.4 and both versions are completely identical. This problem wasn't introduced by one of the updates, it has always been there.

The easiest way to fix this in the level will be to add new links to the "switch stack" of the green square switches. Just keep in mind that Atroxian Realm's "switch stack" system requires a foreground tile with the correct switch attribute for each link in the stack. For the lower switch (the underground one), I would use background tile number 199 (00C7h) and foreground tile number 43 (002Ch) at (X=89, Y=60) and then place the new link on these tiles. That's probably the best combination of existing tiles for that position. Everything else would require new tile images or a bigger redesign of the level.

There are still a couple of other minor issues with the level design that you might want to address if you are going to prepare an update anyway:

- Level 10 (Hornet Hive) should use blocking ceiling tiles at (X=166,167, Y=25) since there is nothing hidden in the ceiling anyway.

- Level 10 also has a some non-blocking ceiling tiles at (X=67, Y=25) that can be used to bypass the lower level on the way back. But since there is only a single Hornet on the lower level, and the players will most likely stun it on the way around to the door leading to the yellow square switch room, there is not much to be gained from this shortcut. This doesn't really bother me, it just seems a little strange and I wanted direct your attention towards it. Just take a look and decide whether to keep it as-is or close off this shortcut.

- Level 12 (Ruins Gateway) has two Spheremints in a secret at (X=31, Y=44,45) that are unreachable because the hole in the wall is only 1 tile high. I put the opening in the wall at 1 tile above the floor level for testing, hoping that this would make sure the Atroxite can't jump into the secret. That seemed to work fine. You definitely shouldn't put the opening directly at the floor level if you don't want the Atroxite to jump into the secret and reveal it (or worse: get stuck in the secret and become a hidden instant death trap).

- Level 22 (Cargo Transit A) has two doors that require a green key gem, but only one green key gem. Since there are bonus items behind both doors, it is impossible to collect all of the bonus items in this level, no matter which of the two exits you take. This could be fixed by adding a second green gem or by moving the bonus items from behind the first green key door ("easy" exit) to an area that can be reached without having to place the green key gem.

If all you want to do is to edit the levels, you can take the "ar.adeps" and "NEW-INFOPLANE.BMP" files from the 1.0 source code release (link), put them into the same folder as the game files from the "unofficial" version 1.4 release and that's it. All I had to do was double-click the "ar.adeps" file and Abiathar had no problems loading the levels. I placed some extra stunners in the BWB level, saved the levels and started the game and everything worked fine. You should have no problems editing the levels.

That's the good thing about how the source code was set up for Atroxian Realm: None of the data files are linked into the executable, you can edit the levels, graphics, sounds, etc. without having to re-compile the code (at least as long as you don't add or remove any graphics or audio chunks).

However, if you want to re-compile the code to update the version number on the startup screen, things will get more complicated. You will need to set up a copy of Borland C++ (3.1) and adjust the directory settings for the project. And because I only included the .OBJ file for the startup screen, you will have to extract the data from that screen before you can edit it and turn it into a new .OBJ file. I can take care of that if you want.

--

Something else to keep in mind is that you shouldn't use your old copy of LModKeen (from the AR 1.0 source release, called modid.exe for some reason) to extract the data from the EGAGRAPH file. The old version will overwrite parts of the sprite image with the red hitbox visualization when you export the "unofficial" 1.4 EGAGRAPH to BMP. That's because I let uGrab optimize the sprites to reduce the in-game memory requirements a bit. You need to use the most recent version of ModId (0.1.3) or uGrab (1.010 or later) to export the images correctly. But your definition file "ar.def" is incompatible with ModId and uGrab, so you need to use this definition instead: And the commands in the argfxin.bat and argfxout.bat files need to be modified as well. Replace "-episode" with "-gamedef" and "-keendir" with "-gamedir" (and replace "modid.exe" with "ugrab.exe" if you want to use uGrab's command-line mode).

The alternative would be to use uGrab's GUI. Simply start uGrab and select "Import" -> "Import from EGAGRAPH..." and open EGAGRAPH.CKD (or a valid definition file, see above), then apply whichever changes you want (select the data item you want to edit and use the Export and Replace buttons). When you're done, select "Export" -> "Export as EGAGRAPH..." from the menu and thave the new file as "EGAGRAPH.CKD" again. When uGrab asks you to select the Huffman compression type, make sure you use either "Keep Current Dictionary" or "None (fastest)". The current configuration of the Atroxian Realm source code doesn't support Huffman compression. The EGAHEAD type needs to use "3-Byte Offsets (default)". You can save a new ModId definition file at the end, but you don't have to.

You can also do a mass export with the menu item "Export" -> "Export to bitmap folder..." if you prefer to go through external files instead of using the uGrab GUI. In that case, you definitely need to save a new ModId definition file if you imported straight from the EGAGRAPH file without using a definition file, otherwise you won't be able to import the files from that bitmap folder later.

Thanks for all this!

If all you want to do is to edit the levels, you can take the "ar.adeps" and "NEW-INFOPLANE.BMP" files from the 1.0 source code release (link), put them into the same folder as the game files from the "unofficial" version 1.4 release and that's it. All I had to do was double-click the "ar.adeps" file and Abiathar had no problems loading the levels.

Yep, worked for me. I had been looking in the 1.4 source code download, which was only the C files.

I just compared the Security Layer level in v1.0 to the one in 1.4 and both versions are completely identical. This problem wasn't introduced by one of the updates, it has always been there.

Apparently indeed. At one point, I'm pretty sure I had a row of foreground tiles with animated links stretching between both bridges, but that probably got broken at some point. The simplest solution here is just to cheat slightly by inverting the rightmost bridge's default state and only have the second green square switch actually target it. It works since you can never see the second bridge before you've opened the first (barring cheating).

- Level 10 also has a some non-blocking ceiling tiles at (X=67, Y=25) that can be used to bypass the lower level on the way back. But since there is only a single Hornet on the lower level, and the players will most likely stun it on the way around to the door leading to the yellow square switch room, there is not much to be gained from this shortcut. This doesn't really bother me, it just seems a little strange and I wanted direct your attention towards it. Just take a look and decide whether to keep it as-is or close off this shortcut

I honestly don't even remember the existence of this shortcut. However, since there was a honeycomb tile over it, my best guess is that it was two-way at one time, and the placement of the honeycomb foreground tile was a mistake that closed it off. I'm just going to shift it over by one to re-open the shortcut as two-way.

- Level 12 (Ruins Gateway) has two Spheremints in a secret at (X=31, Y=44,45) that are unreachable because the hole in the wall is only 1 tile high. I put the opening in the wall at 1 tile above the floor level for testing, hoping that this would make sure the Atroxite can't jump into the secret. That seemed to work fine. You definitely shouldn't put the opening directly at the floor level if you don't want the Atroxite to jump into the secret and reveal it (or worse: get stuck in the secret and become a hidden instant death trap).

Interesting. I've decided to simply remove those points, as there's not enough room to rework that section.

- Level 22 (Cargo Transit A) has two doors that require a green key gem, but only one green key gem. Since there are bonus items behind both doors, it is impossible to collect all of the bonus items in this level, no matter which of the two exits you take. This could be fixed by adding a second green gem or by moving the bonus items from behind the first green key door ("easy" exit) to an area that can be reached without having to place the green key gem.

I've gone ahead and moved those points near the leftmost exit so that they're obtainable without expending the green gem. Since the points remaining window was added in later releases, I think it makes sense to ensure every level is possible to 100% complete. It might not have been the original design intent, but I agree it's better.

However, if you want to re-compile the code to update the version number on the startup screen, things will get more complicated. You will need to set up a copy of Borland C++ (3.1) and adjust the directory settings for the project. And because I only included the .OBJ file for the startup screen, you will have to extract the data from that screen before you can edit it and turn it into a new .OBJ file. I can take care of that if you want.

Yeah, if you'd able to increment it to 1.5, that'd be a big help if it's not too much trouble. I did try to follow the Borland guide a while back to get C compilation working, but I hit a wall at some errors I never figured out how to resolve (sorry, this was a while ago, and I should have made a note of the point in the guide where I got stuck). I have the local changes done, so I could probably drop in the new .exe and prepare a quick release.

To keep this update small and minimize risks, I'm not planning any graphics changes. It's still helpful info though, especially since I've been starting to poke around Omnispeak recently for some AR2 prototyping, and I'd like to figure out the graphics pipeline eventually. In fact, there's a lot I'd like to learn from the AR source as part of the ideas I'm thinking about. I'm able to work with C now, though I'm still in the early stages of learning the Keen codebase specifically.