Development Chronicles

http://m.spectrum.ieee.org/automaton/robotics/diy/youve-never-seen-a-drive-sy...

http://highscalability.com/blog/2011/6/27/tripadvisor-architecture-40m-visito...

photo © 2008 cmcbrown | more info (via: Wylio)

It's been a while since I've posted anything real. I started on a couple articles thinking I didn't have time to update Thumb War, but in the end didn't have time to finish those articles either.

Which brings me to the simple question. Where does the time go? My guess is work and family. I found the above image free online, so I cannot really vouch for it's accuracy, but it sure looks pretty.

On the Thumb War to-do list is:

1. Provide Tablet layouts for those who play on Landscape style devices (Xoom, ASUS Tab, Playbook, HP Slate, etc...)
2. Update to latest AndEngine to get some critical fixes orientation issues
3. Add video of game play to help those who don't "get it".

As for the articles that I've started but didn't finish:

1. Closures: Ruby vs Groovy - Fight!
2. Ghetto Graphs: Too cool for Excel

Enjoy your summer!

Javascript has finally arrived. Ruby watch out!

http://feedproxy.google.com/~r/d0od/~3/aS3EIDQEq6Y/

I've discovered that no amount of screenshots or marketing speak will ever get someone past the $1.99 price point. So with a bit of research and rework a free version Thumb War is now available here.

For a while I thought the process of making 1 project into 2 was going to break the time bank, but after running into this article on the Android Developer site the pieces fit together quite nicely.

The secret is in one property setting of an Eclipse project which allows a project to be shared by another. The setting is "Is Library". Once enabled all classes and resources (drawable, layouts, strings, etc) are available to users of the library. The activities can be addressed in the manifest with absolute package names instead of the relative package names.

The best part of the whole thing is I don't need to create a separate jar build process to make it all work. Now, if you're the type that likes everything automated, I'm leaving that as an exercise to the reader.

After changing the original project a library, I did the following to get the paid version working again.

1. create new Android project in Eclipse with the same package name as my library.
2. edit Android properties and added a reference to the newly converted library project.
3. copied the Android Manifest from the library project into the paid project.
4. changed the activity packages in the paid manifest to use the full absolute package names.
5. removed all drawable in paid project
6. removed hello world in string.xml and updated app_name
7. set svn:externals to the library project's assets directory into the paid projects assets directory, then svn update
8. run

Surprisingly everything just worked.

The next step was to actually create the free project. This is pretty much the same as the paid project so I started with the paid project as the base.

1. copy paid project and rename to lite
2. change package name to com.thethumbwar.lite
3. run

This also worked. It was functionally the same as the paid project and the original pre-library-itized project and both were on my device at the same time. Which means it meets Androids unique package restrictions, too.

The final step in the free project was to add a few restrictions to it and add a few dialogs to encourage the players to purchase the paid version. That fortunately was the trivial work, phew! In case you're curious, I've restricted 3 of the 5 games and if a user selects a restricted game from the menu a dialog will tell them that it's unavailable in the lite and provides a button to go to the Android Market to purchase the full version.

Just updated to the trunk build of AndEngine which broke almost every line of code I had. There were a bunch of renamings that occurred to replace the concept of scene layers with a more generic child entity structure. The shape modifiers also got moved to another package to differentiate from particle modifiers.
All in all, the transition was smooth and I will now be able to contribute to code back into the main source tree without as many conflicts.

I am finding it near impossible to get a perfect simulation of finger to table top physics. I'm not sure if this is a hardware limitation or just my bad algorithm. In a previous post I had a workable solution by dragging pieces around instead of teleporting, and that does work enough. But now that the game is released I feel the need to make is better.
The issue with dragging is the delay that is percieved when moving the object around very quickly. It looks like the object is chasing your finger which breaks the illusion. Fortunately, none of the mini-games use swirling but it bugs me in the Free Play mode.
In the real world, we can put two coins on a table and push one coins into the other. On a touch screen the coordinates come in after the push which is already too late for a perfect sim. So I think some predictive algorithm is in need.
If I can take a previous position and the current I should be able to predict the next and push the object there directly. Sadly, this type of math is my worst suit and I don't have the time to figure it out. If anyone out there lives and breathes math, feel free to comment with a solution. Otherwise the whole word will have to play with drag physics or play a different game. Sad.

Too tired to give any details, but here's a screen for the curious.

I believe I'm almost finished with some thing that I wouldn't hate being in the market. I know that standard seems a bit low, but if I don't release this thing soon it won't ever get released.

The past sessions were all about artwork. This could have been a long rabbit trail, but my wife reminded me to get this thing done. And that every minute I send on this creation, I lose with my son. Anyway, artwork is basically done. I have to crop here and there. But I think it's workable.

Here's some things that are completed:

1. game play of 2.5 mini games (2 full game, and a variation of one of them) 
2. working menus, even replaying a game
3. sound feed back for points in 1.5 of the games
4. theme and art work - most from http://www.lostgarden.com/search/label/free%20game%20graphics who has an amazingly flexible license which I couldn't release this thing without.
5. winning scenes and anxiety inducing count down when game is about to end.

Stuff yet to do before release:

1. add credits scene to fulfill free licensing.
2. add sound to game 1
3. make replay more obvious on winning scene.

Stuff to finish after release:

1. More games
2. More sound effects
3. refactor code base so game 3 is easier
4. respond to user feed back
5. (never ending list)

After release I'll post images of the game...