Our terrain generator is now multithreaded.

I'd not worked with multithreading at all before, having been warned that it would be difficult to learn.  However, while writing and discussing our planned kickstarter pitch (a draft will go up before too long, I hope), I realized that we didn't quite have the tech to emphasize exploration to the degree that we wanted.  We needed multithreading.

So I wrote a new terrain generator - which was much easier than I was afraid it would be - and now it feels like the lid has exploded off the cookie jar of possibilities.  Right now we can generate and instantiate terrain in real time, without any hitching in the Unity renderer.  What does this mean?  It means that we can create infinitely large worlds that expand as the player explores them.  It isn't implemented in a gameplay build yet, but the encapsulated code exists.  This is going to be awesome. :D

E.

Just a few shots of our latest dev build in action.

(download)

Click here to download:

do-a-barrel-roll-ckJclvvtuAyrarBrwEja.zip (2.3 MB)

Rolling isn't just good for dogfighting; it's also fantastic for evading fire that little bit longer when locking on to a ground target.  Just make sure you have enough energy to boost away afterward!

Hello folks,

This post comes a bit later than it should due to the release of Diablo 3.  Oops.

I've added several small mechanics to the game over the past week, including an evasive roll (think Starfox) and an energy system.  Both make a big difference to combat, but the latter is more far-reaching.  Boosting, shooting, or making an evasive roll now costs energy.  Your ship has a sizable store of energy that recharges over time, but the size of that store goes down as you take damage.  This makes combat far less of an all-or-nothing affair; damage matters even if you're not yet dead, and it's more important to put together a plan of attack to get the most out of your energy-based abilities.  Spraying and praying is now a poor idea.  In some ways that makes combat harder, but the addition of the barrel roll is a huge tool in the player's arsenal.  The barrel roll shoves the player to the side and comes with a short invincibility duration during the roll.  This is very powerful for both evasion and positioning, but the energy requirement prevents the player from spamming it.  More abilities and/or pieces of equipment will key off of the energy system in the future.

This upcoming week is a little strange.  I've got a brief presentation to give this weekend as part of the Critical Gaming Project's colloquium at the University of Washington, so some time is going into preparing for that.  I'm currently working on our piece of bigger news though: a kickstarter draft.  It will be modest in scope - we're primarily aiming for a Unity Pro license and a couple of asset packages, which would provide a massive upgrade in Katabatic Ascent's visuals - but it would easily scale upward if interest exceeds our expectations.

Once all that is taken care of, there's another new mechanic that needs to be implemented: takeoff and landing.  The core of Katabatic Ascent is meant to be planning and then executing strike missions, including selecting the right equipment and payload for your planned approach.  That means more pieces of equipment and types of weapons, but before they can be effectively introduced to gameplay the player needs a way to switch them out and re-arm (and the need for a repair system has become clear with the introduction of the energy system).  That means a home base and ways to access it.

E.

Hello again folks,

We're back working on Katabatic Ascent now.  Our 4X roguelike is still something we'd like to do, and will return to someday, but Katabatic Ascent is still our main project for the foreseeable future.  I've been cataloging necessary bugfixes and gameplay tweaks for the past week, implementing a few of them along the way.  There were a lot of basic mechanics introduced in our contest build that we didn't have time to polish at all.  Now is the time to work on that.

Not that we're anywhere near feature-complete, of course.  The whole ship customization system isn't in place yet, and there's a ton of content still necessary to fill in the existing mechanical superstructure.  Yet I'd really like to have something genuinely fun to play in the short term, even if it's limited in scope.  Once a game - any game - is enjoyable, it inspires ideas to improve it.  The imaginative work becomes easier, taking on a force of it's own.  So the goal of this next week is simple: make it fun.

"Fun" is an awfully nebulous thing though.  There's plenty of theory out there on the topic, but the simple idea remains the same: "fun" varies from player to player.  The big issue that jumped out at me this past week was the difficulty of air-to-air combat in Katabatic Ascent.  I grew up on flight simulators, going way back to Chuck Yeager's Air Combat, and I genuinely enjoy the challenges of maneuvering around and lining up attacks based on deflection angles.  It's harder than most arcade combat models, but feels rewarding when your positioning pays off and your target erupts in a fireball.  Not everyone feels the same way though - the small hitboxes, lack of lead indicators, and lack of homing weapons may turn air-to-air combat into a frustrating experience for some.  Only playtesting with a substantial user base can tell for sure.  In the meantime, I'm working on some ideas to reward gunnery and maneuver while providing other players with easier ways to handle airborne enemies.  Those hitboxes are probably going to grow regardless.

Hopefully we'll have some bigger news next week.  Either way, I intend to be much more communicative online in the future.

E. 

It's been a while.  Hello again!

Shortly after our previous post we had a discussion about limits.  The Unity engine is far more capable than our original design of Katabatic Ascent utilized.  We knew this going in, and we thought that it would be fine.  Two problems emerged though.  First, the assets and algorithms we'd created could handle a much more complex environment without much trouble at all - it felt like a shame to not take advantage of that.  Second, it became a surprising amount of work to constrict our algorithms to the limited scope we'd originally chosen; the Unity engine is designed to handle far more than a 2d scrolling shooter, and constraining the engine was actually becoming nearly as much work as expanding upon our core design would be.  So we decided to take a step back and rework the design of Katabatic Ascent.

Our timing turned out to be debatable.  We wanted to submit an entry to the Intel Level Up Game Demo Contest, the deadline for which was last month.  Choosing to redesign our game roughly a month and a half before the deadline was an... experience.  We made the deadline - albeit with a clearly alpha build (which is now available from our downloads page) - but we burnt ourselves out doing it.  For a week afterward, even looking at the code would make me feel nauseated.  To any aspiring designers out there: don't do what we did.  Just don't do it.

That said, I still believe that our redesign is a superior direction for our game.  Katabatic Ascent is now a much more open game, allowing the player to travel freely around a much larger map.  Instead of simply surviving, the objective is to destroy ground targets and avoid or shoot down enemy patrols. There is so much more potential in this design, with relatively little additional work compared to the old design.  Well, little additional work now that we've made the switchover.  Here are a couple of screenshots - it's pretty obvious how different the game is now:

(download)

Click here to download:

redesign-BofcJIDfqiDfpmslkBhg.zip (3.32 MB)

We haven't been inactive since the Level Up deadline though.  Instead, we chose to work on another game while we felt completely burnt out on Katabatic Ascent.  It's something like a turn-based 4X strategy game crossed with a roguelike.  I imagine that we'll talk more about it once it's progressed further - for now, it's been a nice break from Katabatic Ascent.

E.

Hello folks,

Build 037 is up on our downloads page.  It removes the ugly texture seams in the previous version’s terrain.  This turned out to be harder than expected due to some funky interactions between raycasting and the standard way to pause a Unity game.  Still, it worked: terrain shadows are now fully under my control due to manual raycasting, which means that I can avoid the abrupt seams between meshes that Unity’s default vertex shading causes.

(download)

Click here to download:

on-alpha-037-strategic-planning-ecqAgGbkFbdqfIJwjAbb.zip (2.8 MB)

We’ve also reconsidered our plan regarding Katabatic Ascent.  Simply put, we’re not comfortable with trying to sell this game.  Consider how long it’s taken just to get this far; with just the two of us, we’d have to spend an absurd amount of time to reach a threshold of quality that we’d feel acceptable as a professional-level product.  We let ambition and optimism go to our heads in the planning phase.  It’s time to scale things back.  We’re going to build Katabatic Ascent at about 1/3 the size of our original design, and we’re going to keep it free to download and play even after the game is complete.  This is practice for us - pretending otherwise just adds useless stress.  Being realistic about our feature set means we might actually complete a fully functioning, if much smaller, game.  If there is interest after release, internal and external, then (and only then) we’ll plan our how to acquire the necessary resources to develop a professional-grade sequel for sale.

E.

Yup, we're still alive.  Disruptions, both foreseen and unforeseen, have taken a chunk out of us, but we've kept learning and finally have another build available for play.  Build 036, available from our downloads page, adds a simple new beat to the audio and fully textured procedural terrain to the background:

(download)

Click here to download:

on-alpha-036-yfwbtcgtFmxaafgIodqH.zip (1.78 MB)

In practical terms, terrain is my area of expertise.  Back in the day, my main role on modding projects was mapping and texturing.  That was long ago though; some systems have changed.  Now I've been working to modernize my understanding of graphics engines in order to transfer my old handmade terrain skills into procedural format.  That's incredibly frustrating when I can't easily make code do what my brain can reflexively accomplish with Photoshop, but it's also incredibly exciting when I see how much flexibility modern shaders permit.  Build 036's procedural textures are a big step forward, but there's so much room for improvement that it's mind-boggling - and I mean that in a good way!  Mostly.

Next time... we'll see.  Perhaps a change of perspective is in order...

Hello again folks,

I'd like to say a few words on iteration.  Iteration is about repeatedly doing the same thing over and over again, incrementally improving each time.  In game terms, it’s called grinding.  Sometimes you’re grinding your avatar’s numbers upward, sometimes you’re improving your own skill through practice.  I never thought that I’d miss grinding, since I usually despise it within games, but...

Alpha 035 is now online on our downloads page.  It includes basic interface elements as well as more audio experimentation.  We’re getting painfully close to having all core systems present in the game: the overworld is all that remains.  We intend to show a basic implementation of that shortly after New Year’s.  All the elements will finally be there.

Yet having systems “present” doesn’t make a game.  In fact, it’s not even close.  Just because a game is playable doesn’t mean that it’s good.  We need to iterate.  Not just in terms of balance and playtesting, but in terms of our technical execution.  We need to grind our own skill levels in game production.  A problem with implementing one basic system and then shifting over to implement a whole other basic system is that we don’t get much sustained practice working on any one of our core systems.  That will finally change in 2012 after the basic overworld build is done, and for possibly the first time in my entire life I'm looking forward to grinding up.

So obviously the next step is to get cracking on that overworld build.  For me personally, that means the next step is to figure out how to make shaders do what I want in Unity.

E.

Hello folks,

Version 034 is now online on our downloads page.  There are two new, though very basic, features in this version: sound and procedural terrain.  Both have a very long way to go, but it feels good to have them.  Unfortunately the terrain takes long enough to generate that there are noticeable halts in the program when the next level’s terrain is calculated.  That doesn’t matter in the long run, since levels will be separated by a visit to the overworld interface layer, but it’s jarring in the alpha.  Still, it’s been long enough since the last update that I didn’t want to delay further for the sake of an soon-to-be-irrelevant issue.  Here are a couple of screenshots:

(download)

Click here to download:

on-alpha-034-wkCDpBCGBElJubGjfDHu.zip (421 KB)

A little texturing and higher detail would both go a long way, but even this version has some nostalgic value for me.  My introduction to graphical work was the Myth II modding scene back in the late ‘90s.  Myth was an RTS of sorts – real time tactics without base building or reinforcements – that used a mesh system to generate planar terrain.  I wrote a similar algorithm for the procedural terrain in Katabatic Ascent (KBA).  Unity is capable of much more though; it doesn’t matter greatly for KBA, but Unity doesn’t restrict me to a simple plane of points with varying heights.  Things like procedural caves under the procedural surface are entirely possible in Unity; more complex to calculate, but entirely possible.  So much potential for the future…

The overworld and logistics system are now the only remaining untouched elements of our design.  I think we’ll spend some time improving what we’ve already got, but I’ll want to start implementing the overworld as soon as we can.  It will be much clearer what sort of gameplay experience we’re ultimately aiming to provide once the overworld is playable.

E.