2000 January 20 - 00:00 Old batch converted post
2010-04-08 UMAG Released
UMAG version 1.0 is finally released! :)play at candystand
— A short post mortem or summary of the development —
Early Development
I had developed several multi-player games before but all of them where based on existing platforms, this time I wanted to create my own.
The idea for UMAG was just to create a multi-player game as simple as possible to learn from and create my own multi-player server.
Making a basic artillery game seemed like one of the most simple yet fun multi-player games.
Alpha launch
The first evening when Unnamed Multiplayer Artillery Game open alpha was released on Kongregate pretty much nothing worked, there was numerous bugs, mainly on the server, one of the more serious ones caused the cue for games to break and there where hour long waits to start a game.
For some reason people still seemed to like the game and kept trying to play, kong rating was at about 3.40.
After about a week and some frustrating bug fixing and learning Java better the game got fairly stable, and both players and rating started to rise.
FGL
I was a little worried that having most of the game already available on Kongregate might discourage sponsors, but I think it went the other way, since the alpha had such great rating on Kongregate (something like 4.30) it was already proved to be a success.
Tools, tech and stuff
Game client is written in AS3 and server in Java, using MySQL for database and is running on a debian server.
I use the Eclipse plug-in FDTFDT for the flash development, I compile with the flex sdk and try and use Flash CS to a minimum, I hate it >_<
For Java I use Eclipse too, and for the UMAG server I used the framework Apache MINA to handle Javas nio-sockets.
For graphics I(we) use photoshop, my friend Jonas Johansson made most of the graphics.
What went right
Public testing = win
Testing mutliplayer games by yourself is just really hard and boring, also most of the early bugs on the server only surfaced when having more than 30 simultaneous users, it was really good to find and fix these problems early on, and not having an nightmare late in the development.
The quality on most of the bug rapports where of course very low, but that combined with a lot of logging helped tremendously with the development.
At the time of release at candystand.com the combined playtime was just over 39 years, which I would think is one of the most massive beta testings ever made for a flash game. ;)
The game got way more popular than I had expected and is currently the highest rated multi-player game on Kongregate even though the final game is still exclusive to Candystand.
What went wrong
When starting the project I just wanted to create a simple multiplayer game and get it out quick.
Therefore the base of the game was more sloppy than it should have been, also I kept all the stats saved as flash cookie and used mochimedia’s leaderboards.
However as the game became more popular and competetive it got indefensible to keep the stats and leaderboards on the clients.
I had to rewrite parts of the game to make it handle more on the server and make it more secure.
Desync bugs.
I had problems with clients getting slightly out of sync since day one, sometimes, although pretty seldom and never possible to reprocude locally, the tracetories on one of the client in a game becomes slightly different, causing different results on the clients.
Eventually I made the clients take screenshots of the game and send to my server, I collected huge amount of data but never manage to fix the problem entierly.
Abusive users.
There’s a shooking amount of idiots at Kongregate, players who behave really badly, spamming, calling people names and harassing them. I should have implemented a ban och mute system from the beginning.
What’s next
There’s probably still a few patches to be made on UMAG, and the viral version is not yet released.
I’m probably going to filter the chat, and ban abusive users.
I think multi-player games it the way too go for me, there’s such a huge amount of singleplayer flash games out there that it’s really hard to stick out and get noticed.
There really isnt a lot of mp games, a mp game sure requires way more work than a single-player game but due to it’s rarety and replayability also attracts more and by far more loyal players .
Many UMAG players have played thousands of matches and still spends hours daily playingview leaderboards it.