I've been working on lol-wars for the mobile for a couple of days now. The gameplay is pretty much the same as the original html5 version, but the experience is what i've been working on.
The major problem with the html5 version was the use of firebase for handling the communication between the players and also to keep track of who is online. Frankly the code was all messed up and all over the place. I did however learnt how to use angular material and get familiar with the riot's developer api.
For the mobile version i've tried doing it with unity, quickly realising that i was spending way too much time without much progress. I was trying to figure out the gameplay flow while trying to figure how it's done and eventually the project got sidelined.
I still loved the idea but wasn't able to motivate myself to get it done in unity, so i thought to myself that getting the game done was more important to me than doing it in a particular tech. I needed this game to be available on mobiles. So i picked cocos again and this time i did the same mistake of trying to pick up new tech .. i.e the server side programming. I wanted to use netty so i get a good understanding of how its' done etc, but was too frustrated trying to set the machine up. I discovered a library for socket.io which is fantastic because some of the basic stuff required for a turn-based game are already provided, The best part?. cocos2dx has support for it..
After fiddling around for a couple of days i got comfortable with the Socket.io client in cocos, but the server was another beast. I finally reverted to nodejs. I initially wanted to try it out but backed out because of perf concerns but not any more.. the server can always change.. but i needed something working. So i went ahead and implemented the server in nodejs using the socket.io library.
I was quickly able to get the protocol decided for the game to work. I also quickly did a couple of state animations using cocosbuilder and got the screens ready for gameplay. Over the weekend i was able to find enough time to finally get the server and client up and running.
The problem now was to figure out how to host this server online so i can access it anywhere. I looked at a couple of options...
1) RedHat redshift
2) Windows Azure
3) AWS
I've spent about half a day trying to host the damn app on RedHat redshift service, but with no success. I mean i was able to get it running but was unable to connect to the server with cocos2d-x client. I didn't have the time / patience to figure out why. I tried looking for quick solutions but nope nothing helped. The server was up, the browser was able to connect to it but not the cocos2d-x client which kinda makes it worthless.
The windows Azure asked me for my credit card info and when i was checking it, i was confident that i would be able to get the RedShift thing work for me. But finally i had to give and pick the AWS Free Tiere option.
The best part of this decision was that i finished the setup and testing of the whole thing in about 20-30 mins time. Which is incredible considering how much time i spent trying to figure what was going on with rhcloud.com
Any way just happy that i was able to get this running. Next stop getting more polish in, and trying to see if we can make this look and feel better.
The major problem with the html5 version was the use of firebase for handling the communication between the players and also to keep track of who is online. Frankly the code was all messed up and all over the place. I did however learnt how to use angular material and get familiar with the riot's developer api.
For the mobile version i've tried doing it with unity, quickly realising that i was spending way too much time without much progress. I was trying to figure out the gameplay flow while trying to figure how it's done and eventually the project got sidelined.
I still loved the idea but wasn't able to motivate myself to get it done in unity, so i thought to myself that getting the game done was more important to me than doing it in a particular tech. I needed this game to be available on mobiles. So i picked cocos again and this time i did the same mistake of trying to pick up new tech .. i.e the server side programming. I wanted to use netty so i get a good understanding of how its' done etc, but was too frustrated trying to set the machine up. I discovered a library for socket.io which is fantastic because some of the basic stuff required for a turn-based game are already provided, The best part?. cocos2dx has support for it..
After fiddling around for a couple of days i got comfortable with the Socket.io client in cocos, but the server was another beast. I finally reverted to nodejs. I initially wanted to try it out but backed out because of perf concerns but not any more.. the server can always change.. but i needed something working. So i went ahead and implemented the server in nodejs using the socket.io library.
I was quickly able to get the protocol decided for the game to work. I also quickly did a couple of state animations using cocosbuilder and got the screens ready for gameplay. Over the weekend i was able to find enough time to finally get the server and client up and running.
The problem now was to figure out how to host this server online so i can access it anywhere. I looked at a couple of options...
1) RedHat redshift
2) Windows Azure
3) AWS
I've spent about half a day trying to host the damn app on RedHat redshift service, but with no success. I mean i was able to get it running but was unable to connect to the server with cocos2d-x client. I didn't have the time / patience to figure out why. I tried looking for quick solutions but nope nothing helped. The server was up, the browser was able to connect to it but not the cocos2d-x client which kinda makes it worthless.
The windows Azure asked me for my credit card info and when i was checking it, i was confident that i would be able to get the RedShift thing work for me. But finally i had to give and pick the AWS Free Tiere option.
The best part of this decision was that i finished the setup and testing of the whole thing in about 20-30 mins time. Which is incredible considering how much time i spent trying to figure what was going on with rhcloud.com
Any way just happy that i was able to get this running. Next stop getting more polish in, and trying to see if we can make this look and feel better.
Comments