Some of my creepy crawlies are walking into cracks and getting stuck there. With no obvious solution, I’m taking a programming break and reading up on marketing. The consensus among published game developers is that developing the best game in existence isn’t good enough–you also have to market it really well. Here is some of what I have learned.

First of all, my friend Cory recommended The Big List Of Indie Game Marketing, which is a collection of links that go into great depth about every aspect of indie game marketing. If you want to learn this stuff, know that you can’t absorb it all at once, so be patient and just start reading.

A lot of what I have read–on that list and other places–indicates that the most important marketing tools for an indie game developer are a website and a blog. This website has existed in one form or another for around fifteen years–seriously, has it been that long?–so my basic web presence is taken care of. Web traffic is consistent too, with an occasional huge spike when some big-name blogger like Chris posts something about my site. I added the blog last June to try to increase traffic, build an audience for Retrobooster, and get constructive feedback from anyone who wants to give it.

I have also been working to improve other aspects of this website recently, putting a Retrobooster level playthrough video on the front page, creating a page devoted to Retrobooster to give some details about the game, and adding social sharing buttons (Like, Tweet, +1, etc.) to some pages and each blog post. (It turns out a lot of people had already Liked and Tweeted my site even without those buttons. Isn’t the Internet neat?) I also removed the donations page from the screensavers section. The donations have been great and I appreciate each one; they allowed me to get a new graphics card a couple months ago when my old one died. But now I feel a bit weird having a donations page when I’m trying to transition the website from only giving away free stuff to actually making money. Similarly, I removed the Google ads from the front page while leaving them on several other pages because I had read some comments online that my ads on the front page make the whole site look a bit sketchy and untrustworthy.

There is still plenty to work on, like needing to figure out how to use all this new web infrastructure to build hype. And a lot of smaller things like beautifying the social sharing buttons, which are way too bright on this predominantly black website.

The bottom line is that marketing is about communication. I need to do a whole lot more of it in smart ways, such as by getting interviews, previews, and reviews just as soon as I am ready for that stuff (so far it seems like nobody wants to do an interview until you are about ready to release your game), maybe building a discussion forum, etc. It would be great to already know how to do marketing properly, but this is a good learning experience. Please let me know what you think about all this in the comments. Or send me an email if you want to talk about marketing or anything else.

I have decided to make Retrobooster my full time job soon. Reception among testers and the Internet community has been favorable, and it’s the most fun art project I’ve ever had. It will be great to be able to concentrate all my efforts on this game. First, I have to wind things down at my day job, which should take a few weeks. I’ll post another announcement here when that’s done and I get cracking full time.

When I started development on Retrobooster, the main design goal was to make it feel like a big toy you can play with. Later on I read Jesse Schell‘s wonderful The Art of Game Design and learned that he had already named this idea The Lens of the Toy.

The best way I have found to make Retrobooster toy-like is to make each level a unique toy. Testers have been letting me know which levels keep their attention the best, and I have a chart in my notes that lists what new concepts I introduce on each level. Testers pretty clearly start to lose interest on the levels where I have neglected to introduce something new and fun.

Lately, I have been working on a level that is an homage to Asteroids, which is a good example of a toy level. It’s hard to imagine making this game and not throwing in some asteroids. After all, it already has a thrust ship and some decent physics for bouncing things like asteroid off of one another. Of course, I had to add Retrobooster’s sense of fun by making alien critters attack you while walking around on the asteroids.

There are other, more subtle ways that make this game like a big toy. Once you unlock a level, you can start a new game on that level anytime you want. This is more like something you would see in a racing game instead of a linear action game, but it supports the toy concept really well.

I suspect the biggest deficiency right now is that the weapons don’t feel enough like toys. Testers spend most of the time using the primary weapon and often forget about the secondary weapons. Part of this is probably because the primary weapon is too powerful right now, but part may be because the weapons just aren’t enough fun to use. I’ll have to keep working on this and tune the weapons a lot better.

Programming lightning was way too much fun, so it would be silly to not give it a blog entry. Plenty of games have lightning, but the style and quality varies a lot. I wanted mine to be as realistic as possible. Later on, if I don’t think it fits with the style of the game, I can get more artistic with it.

It took a fair amount of fiddling to settle on the best way to render lightning. To make the basic lightning shape I start with a segment between the start and end points, divide it in half to produce two new segments, and displace the center point by a random amount. This process is repeated on new segments until the segments are small enough to make the lightning bolt pleasantly jagged. Smaller bolts are allowed to stretch out from some of these points, much like the way real lightning branches.

In a similar manner, a velocity is assigned to each point. The two endpoints are assigned velocities based on the movements of the entities connected by the bolt, and the in-between points have averaged and then slightly randomized velocities. This makes the bolt appears to warp in random ways before it vanishes. The movement is subtle, but it makes the bolts look much more natural and alive.

All these points are connected by thin polygons that face out of the screen. A simple texture is applied to the polygons, which makes the bolts look fairly realistic and much less polygonal. The textured polygons are drawn as pure white, and their brightness flickers a little.

To add color I draw a colored version of the same lightning bolt into a glow buffer. The glow buffer is then blurred and drawn on top of the existing image to make lightning and other specific parts of the scene look brighter. This creates a colored aura around the white lightning. Color is also implied by light sources that illuminate the scene near the lightning bolts. It would have been difficult to make decent lightning without already having good lighting and glow systems.

A good lightning sound effect took a while to find. The usual tzzzz tzzzz electric sounds are alright for short circuit effects, but they really don’t sound like lightning. I wanted loud pops like you would hear from a large electrostatic generator. The solution I found was to use a truncated bullwhip sound. If you skip the whoosh leading up to a bullwhip crack and play the sound starting in the middle of the crack, it sounds quite electric. I also add some rumbling thunder on occasion for good measure. You can hear all this in the accompanying video.

This week I nerded out pretty hard on making panoramas from exposure fused images. Exposure fusion is something you might want to try if you have a scene with great variation in brightness. By taking pictures at different exposures you can get images that show detail in the bright areas and others that show detail in the dark areas and then blend them together. Exposure fusion is a cousin of high dynamic range (HDR) imaging. However, instead of creating an HDR image and tone mapping it down to an LDR image, you get a relatively realistic looking LDR image with less hassle.

The city lights scene above makes a nice panorama but doesn’t show off exposure fusion particularly well. The image of Hangar One at dawn has a good variety of light and dark areas and is a much better subject for exposure fusion.

To make these images I used Hugin, which is an amazing panorama stitcher that also does a great job of exposure fusion. In Linux, just install hugin and autopano-sift-C, which is the default program for creating control points in Hugin. Sometimes it blows my mind how much awesomeness you can get from free software.

This week the 4th alpha release made it out. It is the first release that I have been able to package and send to testers to use at home. It was an important milestone because there is now too much content in the game to do a thorough in-house test. In the future, I will have to test specific pieces of the game in-house and let testers go through it more thoroughly at home.

Good observations and ideas have already started to come back from testers. This will help drive the next milestone, which will focus more on content creation. Retrobooster needs more and better levels and enemies. There are also plenty of things to improve in the interfaces.

As a bonus, I was able to send Retrobooster to Kevin Meinert of Subatomicglue. He has agreed to work on music for the game. I already have some of his older tracks in the game and testers often think they were made especially for it, so his music seems to be a good fit.