New home, new start

The past half year has been really enjoyable personally, but bad for my side project. My wife (as of August) and I left our home in Boston and travelled for four months through Europe. Then we were in Ontario for Christmas, and now finally we have settled in our new jobs here in Minden Nevada. For those who need a brushing up on American geography, Minden is an hour south of Reno, just across the border from South Lake Tahoe California.

I won't lie, my odometer project stalled. The Freescale kit got packed before the trip and is currently sitting somewhere in storage here in Carson City, waiting for us to take possession of our new home. I do have some positive news regarding my exploration project. During the trip I made an effort to put some time in to it and did get a python project to about the same level as I had with java. The fruits of my labour can be seen below:

Yes, those who are a little more astute will notice that I am now developing on a Mac. My wife is an avid Mac user and we got a macbook air for our trip. Python has the beauty of being cross platform so I could easily take the same source code and bring it to my PC (which is also sitting in the same mysterious storage locker as my kwikstick kit).

I did not do any bound checking on my project so if the boat goes too far north or south it just floats off the map. This program DOES load a map from a text file and draws land vs water which is a step further ahead than my last Java attempt. I am moving towards an MCV (model controller view) model of programming. However, the project currently uses an event manager which I am having second thoughts about. On an airplane ride with my wife I talked over my problem and collectively we came to the conclusion that rather than having a 'dumb' controller I should have it do the bulk of the model manipulation. This would fit more cleanly into the MCV paradigm and allow me to change the interface without affecting game mechanics. I should also be able to easily jack in AIs for the computer players. Fun stuff.

Stripping out the event manager and adding smarts to the controller are my next steps. It will no doubt take my progress back a little. But let's call a spade a spade, my project is not off the ground yet.

Until we meet again...

Epic Fail

The decision to stop developing with Java was likely best for long term usefulness but it was immensely painful on the immediate schedule. As you might have guessed, I have not been doing any development on my game in the past few months. That is not to say I haven't been programming, just not high level programming.

The new project I've picked up and have slowly started putting hours into is using the Freescale Kwikstik to act as an odometer for my bike trainer. I am hoping that by using the onboard IR diode and receiver I can detect a piece of black electrical tape spinning around the inertial mass. In my mind I should be able to pull this off quite simply and then once I'm counting revolutions the project can really take off, USB data uploads, speed profiles, etc.

If I do get data up to the PC it would be in the realm of my fiancé Jackie to decide how to layout and display information on the screen, I imagine I'd write the interface though I'd be hacking my way through it. That slice of software remains the missing piece between our skill sets. I know that we both talk about wanting to learn it, we would be quite a team if we could meet in the middle. Me from the drivers up, and her from the screen down.

Currently I have the LCD screen driver, and the IR diode on, my immediate plan is to get the IR receiver working. I've discovered how difficult it is to work with embedded systems without the aid of meters and scopes. If I get serious about this hobby I should look into getting some tools.

Competitive Analysis

I recently purchased Commander: Conquest of the Americas, through Steam, as it looked like it fit into the genre I'm hoping to enter.  My intentions were to evaluate this game and hopefully incorporate the best aspects of it.


Let me begin by saying that this game was not well reviewed.  It wasn't poorly reviewed, just deemed mediocre.  After my first couple hours in it I can see why.  It lacks the intuitive simplicity that Sid Meier's games are famous for.  Though that is a hard standard to compare against.  The team at Firaxis always amaze me with their balance of simplicity and strategy. 

Commander seems to be a trading simulation at its heart and in that respect I compare it to the Patrician III.  In fact, I would say that it is a step closer to my intended game than Patrician is.  You found colonies, bring colonists over from the Old World, and compete with fellow European Nations.  The focus though is on developing the colonies production trees, and setting up automatic trade routes that maximize profits.

I haven't played through a whole game yet so my summary may be premature, but my first impression take aways from this game are as follows:

GOOD

• Resources distributed around map, semi-realistically
• Not based on a grid, ships can move around in any direction similar to Pirates!
• Resource development, you can improve colonies to turn local sugar cane, into sugar, and then rum
• Ships can easily form squadrons and fleets to make protection and resource pooling simpler
• Four advisers that give suggestions and missions to complete
• Very pretty to look at
• Steam achievement system

BAD

• No exploration, the whole map is known from the start
• No randomization, you always play the map of the Caribbean
• Too many resources and colony improvements
• You must keep the advisers happy or the game will end
• UI is bad
• No homeport screen, colony screen hard to access and provides little information
• No good statistics or summary screens, information is hard to find
• Tutorials are just videos with someone speaking about what they are doing
• No end game, the goal is to amass a fortune

 I'd say the game has been worth the $15.  It provides another reference point for me, and there are some aspects I did enjoy and will try to build in.  I don't know if I'll go straight for the 3D rendered ships and oceans, but the original Exploration could do with a more advanced economic model.

Was Java the right choice?

As I started to expand from a single ship navigating a 7x7 map to any number of ships navigating a nxm map I started to add more and more Java classes.  While this in and of itself isn't so scary, the thought occurred to me that in order to eventually do the port to the iPhone, all of these java commands and classes will disappear, leaving me with gaping holes all over my program. 

The past couple months of working in Java has given me a better feel for the strengths and weaknesses of this language.  It is strong because of the documentation and huge community.  It also has implemented a lot of functionality that is simple to import and use but therein lies my discomfort.  Java is meant to be a big black box.  You write code to make one black box work with another and I don't like that. 

The decision I have to make now is do I continue to use these boxes, or do I transition to C++ and leave java behind.  That is the way I'm leaning.  I might start a parallel project to see how long it takes me to recreate the functionality I have now but without all the black boxes as my foundation.

Approaching 0.1

Since my last post I've made good progress.  I found a better ship image, drew an ocean tile that looks better than straight dark blue, and I can move the ship around discretely by squares.  Currently the ship can move around a 7x7 grid.  When he hits one of the poles he can't go any further, and when he goes off the end of the screen, he appears on the other side.

To get to where I hoped to be at 0.1 I need to implement the following:

• Expand the map to a size greater than what I can display in a 7x7 window
• Keep the screen centred on the ship unless he approaches a pole
• Add greyed out unexplored squares that are revealed when the ship is next to them

So that should keep me busy for the next week or two, hopefully I'll also layout the project development milestones so I can break up the project into manageable chunks.  Currently I find myself wanting to work on every feature at once and as a result spend a lot of time finishing very little.

Ship ahoy!

In light of yesterday's realization that I need to spend less time paralyzed by indecision over how to structure my code I went home last night with one goal in mind.  "I will draw a ship on the ocean, dammit".  I was fairly close already, having adapted a lot of the window opening and image loading from the Space Invaders tutorial.  I found a ship image online and focused on displaying it.

It took me a little more than an hour during which my fiancee correctly identified that I was paying less than full attention to any attempts at conversation.  (Sorry!)  But before I went to bed I did indeed prevail.

The following is a screenshot of a ship on the ocean-ish. 

One small step...

Found another great tutorial.  Rather than copying it verbatim I'm using it as an example.  It is a recreation of space invaders and can be found here.  They do some interesting things with resource management and I've learned about learned a fair bit about loading images into Java. 

Also, I found an interesting article about game loops and screen refreshing.  That article is here.  The original exploration had very little animation, and a very blocky refresh rate.  It will be simple enough to recreate, and not much harder to smooth out.

As with most ventures I find I keep hesitating during startup.  I keep falling back to doing more research into Java game programming instead of moving forward with my best guess.  I need to remind myself there is no cost associated with failure and I can always go back and optimize later. 

Currently I'm opening up a window, but not putting anything on the screen.