So, to make things easiers, here goes the download link for my sketches of the chapter screens. The finale is yet to come and if it comes, I will transfer it privately to the artist (if its not a matter of me doing it).

http://www.easy-share.com/1909704627/lifeonlies_chapters.zip
Things to keep in mind:
- No gradients
- No textures
- No Curves
- No outlines (yet)
- No more than 128 vertices per polygon
- Alpha blending works very well
- Freehand works, but use it wisely.
- Never tried anything beyond inkscape
- Never tried embedding bitmaps inside the SVG, but probably wont work.
In case you want to know how something looks, take Droid Hunter and start modifying the “chapter” files.

In case anyone wants a pre-release version of the episode, let me know.

Im still amazed with how slow can Windows Mobile be with graphics…

I tried setting a new MAEMO-26 icon in the control file and due to conflicting documentation I created bigger icon. The problem is that dpkg doesnt know how to handle it and it probably results in a buffer overrun. dpkg fails and leaves the system in a inconsistent state. The sad fact about it is that dpkg says your file has a unexpected end – exactly the oposite!

Maybe this is actually Debian’s fault.

In case you happen to see this happening, just try to fix your MAEMO-26 icon.

- You cant use dirty rects, you got to update the whole thing.Still, you need to use SDL_UpdateRect instead of SDL_Flip

- Even as you will only update 0,0,400,480 , you have to set the surface to 800×480, otherwise, things will get nasty (and belive me, I got it a lot until I figured it out)

- remember to “-lXsp -I/usr/X11R6/include -lX11 -L/usr/X11R6/lib “

- dont draw outside the window, or it will flicker back to 800×480 and then back to 400×240 (other than that, no problem)

- graphics will look granier, but its not the end of the world. things will get slightly a lot faster.

I wonder if this works for N8x0… It does work with N8x0 too. And its even faster!

The Genesis:

As soon as I finally had BZK running on my cellphone, I realised there were some demand for games to fill this niche. Sure we had Quake and everything. But as much as I had played though some 4 levels or so, soon it got just too dificult to play. Its not meant to be played on a phone.

Just as I was to finish The Riddler, I was greatly disapointed with it. While the levels looked good, the puzzle factor was not good enough and its speed on the mobile devices were not al that great. No complaints really – it was meant for the EeePC. This disapointment increased when I reached the dreaded last 10% and got myself unmotivated to finish it.

On the other hand, I was very curious to see if I could make a good action game out its engine. If Angstron made me finish a usable version of it, The Riddler made me greatly expand, improve and optimize it. It was time to make use of the features left out. And as a last addition, I wanted to finally implement a simple scripting scheme.

First steps:

A few weeks before the Global Game Jam 09, I did some small tests on what would be Angstron2. Initially a very diferent game – much more related to Angstron 1. A lot more like Forsaken or Decent , but with walking levels too. But the idea was quickly scrapped, since it was not fast nor easy enough to design or to randomly generate good levels. On the other hand, just as the engine was very good on indoor enviroments, I tried a diferent concept and it kind of worked. I had some leftovers from angstron 1 and riddler in a closed enviroment and some simple first person shooter logic in place.

EDIT: Now I realized this is probably the oldest image containing Droid Hunter elements. Doesnt look like the fist level?

Proper execution:

My first steps on this was to finally ditch the C++ implementation of the game logic and implement every table and FSM state I needed to get a equivalent behaviour. It was rather quick, since I had it sketched for almost a year already.

Then it was the time to finally make some use of MonoBED. Droid Hunter was the first and the only game entirely designed in it. Using it was rather quick and easy, but somewhat laborious, since the BSP compiler was something new and I didnt really knew all its properties. I had to manually allign everything as I didnt knew the compiled handled this for me.

After the rough laying of sectors, I had a few adjustments done to improve speed and simplify design and gameplay. Until then, only 2 months had passed since the project started and things looked bright. And  then, I was broke and needed a job.

I finally god a job as a sys admin at univerisity. Nothing really that bad – 20 or so huge motherfucker computer stations running 64 bit Linux at my disposal. Needless to say I had plenty of time to work on my projects. No wonder MonoBED blossomed so much at this time.

Tragedy:

One day my  EeePC misbehaved and I decided I needed to re-install the Xandros from the CDs. The problem is that the CDs doesnt state that it will wipe your HDD to install it. I was happily wiping out everything and my backups were far from complete and recent. To Angstron2, this was a big setback, but not as big as Riddler – it was shelved.

The recovery was bitter: I had to work post-processed levels and with little to no help from the level editor. Luckly, BZK uses XML and I could write myself a couple of tools to help me working around it. At this point, it was 3 months of project. Only 1 left from the planned 4 months.

The last month was mostly spent polishing the engine, the FSM and placing objects and textures. Also, during this last month, both Level 2 and Level 3, one at the time, was saved from the axe. Initially , Level 3 was very problematic (mostly due to bad design choices) and very slow. Also, the planned gameplay was dificult and was severly simplified. Then it was Level 2. While it was very simplistic and in some parts, very fast, it was at first, somewhat buggy and had some very specific slowdowns (one still not solved – and wont be, since its not crictical)

Faux Podium:

4 Months later, the game was finished. But it was far from satisfactory. Even now, little have changed in terms of gameplay, but graphicly the game looked poor and a rushed piece of work. But I wasnt interested in working on it. I was about to take a new internship…finally in a video games company! Its called Nano Games and it produces mobile games. In my job interview, a unfinished Droid Hunter was the central piece of my portfolio. I showed them the game running on both the EeePC and the A1200i and I belive it was a good display.

The idea of finally working on the RPG Im designing since high school seemed apropriate. Specially since Im already making games taking place in its universe. I decided to go for a semi-sequel for The Riddler and making use of the guns initially designed for Droid Hunter.

But then, I got myself on the Festival de Jogos Independentes (Independent Game Festival) for SBGAMES 2009 (kind of a brazillian GDC) and I decided to tweak Droid Hunter for it. The submited version was mostly a rework of the 2D elements, like the textures, cut scenes and the gun, but the idea improving the game play and fixing more bugs was soon born out of the enthusiasm of the improved version. It finally looked good!

Unfortunately I lost the IGF (I belive I lost it to the iPhone hype), but I was entoxicated by the idea of how the game could get. I planned a big release, full a great new features. And along the way, I applied for the GPC08, the contest of games made in 2008/2009 from people of the PDJ community. And now I can say that it was worth it. I won!

With the new features, I finally had TCP/IP multiplataform multiplayer (albeit, only local, two player games are supported), better resource handling even more graphical and gameplay tweaks. The game finally got most of its final look at this stage, as I wanted to show it to my friend at the SBGAMES09 and wanted them to test it. The insights were very helpful indeed. One last work sprint had to be done.

The final countdown:

One last sprint, as I finally was in vacation from college and only had to go to Nano Games to work, was possible to manage. During this last sprint, I also had the chance to kickstart something I always tried and never quite achieved: a Java Mobile version. Now, thanks to the JSR 18 4 (a.k.a. M3G), I finally have a simple demo that shows some good future ahead.

This last sprint, finishing this weekend was very simple and was much more to fix small bugs and make room for the newer episodes. I also must thank James B “supertaco” Lynch and the folks at Linux Game Tome for supporting the project. James, in special, was very helpful and decided to remake the Vacão model. Replacing it was a much cooler Gargoyle that you see now in levels 3 and 4. Unfortunatly, he also did a new Jane model that I dont have the proper time to animate and use – but its saved here and will eventually be integrated in the new episodes.

The Future:

In my first emails to James I promissed him a Android version of Angstron 2 and I intend to keep my promisse. The main problem right now is that it seems that the native interface doesnt support STL. Since BZK was designed to handle those cases, its just a matter of reimplementing the container interfaces – but this will probably happen after I implement a OpenGL ES version.

About a month ago I also purchased a Dell Axim x50V from a brazillian auction site and this allowed me to restore the Windows Mobile port – but its very slow to be played. The OpenGL ES version will be faster.

I also got the obvious conclusion that supporting the Xandrois EeePC port is a waste of time. Nobody uses the stock xandros anymore. But I want it to still play it, so the desktop versions will all be compiled in it, since libc5 apps run on libc6 too.

But I belive the most obvious step from now on is the renewed focus on Maemo and Windows Mobile. Besides the fact that right now the Nokia N770 sits on the lower end of chain devices I have here to debug my games, its a great platform (still kicking ass, after 5 years!) and much more important than fluid framerate and realistic scene, the social aspect of the plataform allows for some great possibilities of gameplay. Of course, the maps are still tested on it, to make sure they are at least playable and the engine will still be tweaked for it.

And now, for what most were expecting:

What went right:

Graphical Style:

From the colors and lightning, to the 2D art, everything had its own style. It doesnt matter if it looked like early 90s graphics and wasnt as nearly fast as quake – it had its style. No wonder some took it for cell shading. I was VERY LUCKY to choose the right colors and as a result, some levels do really look good. Im afraid indeed of how it will look on more resourceful OpenGL ES ports, with regular textures and lightning.

Simple gameplay:

As a testament of this early design decision, I usually like to play Droid Hunter with only one hand at the ferry, when commuting. The ideal of a one hand game was great, since I dont have to ackwardly hold a stylus and scratch my screen to aim on a tiny LCD. We cant think of FPSs being the same for mobile devices as for desktops. The only mobile FPS I remember being TRULLY GOOD was Ashen on the original N-Gage. That game kicked Assh.

Simple and effective sound design:

I wanted something better, and a better theme song indeed – but was out of luck on this. While the theme song for Angstron was very good, I cant say the same for Angstron2. It was such a gloomy sound that I decided to save it for some special moments into the game. Other than that, while being a rather silent game (great for playing on the go), the game is effective on setting the mood for levels. The SFX match the comics quite well and the sound effects were greatly revamped at the second sprint.

Project size:

keeping it simple and small was a good idea in many ways. Simpler means easier to develop and easier to grasp. Also, this also meant less data to mantain and less art being needed. Also, the rejected elements now have to chance to mature and return on future episodes, finally fitting into the project universe.

Vector graphics:

The very nature of vector graphics make it the best for mobile games, since there are no standard screen resolutions. The ability of scaling it and not worrying about it is priceless.

What went wrong:

Organization:

Somewhere in the third month of project, I failed to backup my work and lost the original set of levels, only leaving me with the post-processed levels stored inside the memory card of my phone, to be tested. If I had the original files, I could place objects onto the maps much easier. Maybe the project would take 2/3 of what it took.

Focus:

What went well on the first 2 sprints was that I had a good list of things to do and knew what to do. I was focused and finished it rather quickly. The problem was then the famous last 10%: I wasnt satisfected and didnt know what could help. Maybe I still will see room for improvements, but now its better to let it go. Again, this is much more a organizational thing, but I having college+work+family+girlfriend is not easy. As much as I own portable computers, I live in Rio de Janeiro and its not very safe to carry it around. Sure, I developed the FSM at my gf’s house, while she was sleeping and the multiplayer finally got to work when I was coding in the ferry, but coding on the EeePC for production code is not very confortable.

At the later dates, the offsprings of the main project also got in the way. The M3G and the OpenGL ES versions trully delayed the last sprint. But I see this as something good and that might payoff in the near future indeed.

Concentration of tasks:

Is not hard to see that a programmer making art will never be that good. Please, dont stare for too long on the chapter two comics. Luckly, I got James on board this last month and while he’s not full time into it, he does help me quite often. Now I just need some more 3 or 4 clones of James to get a great game.

Vector graphics:

Sometimes, I feel that what takes me weeks to achieve with vectors would take hours to make with bitmaps. But Im constantly working on the constraints here. Otherwise it would be too easy and not very interesting.

Overall impression: Very positive. I feel Angstron2 as a full blown game, not as technology demo, just as the fist Angstron. Its a matter of learning the patterns and try to repeat it.

Droid Hunter’s sequel coming very soon!

And maybe PocketPC too…