I decided to reset my blog and transform it to a part of a micro-blogging system using ping.fm. Stay tuned.
Posts from April 2009
testetseb alb alb lablaaa
I’m using GlowCode trail version for profiling the game. i have to say it is a fantastic tool. I managed to find the performance bottle-necks in 5 minutes and fix them. Link-Dead is running super smooth now and I still got stuff to optimize.
Since Easter Holidays is all about killing and torture I decided to give you a little present. Here is an actual gameplay video from Link-Dead with bots. The graphics and animations in the video ARE NOT Link-Dead graphics, they are placeholder graphics I made in 5 minutes. I know there are a lot of retarded people out there, so please remember this and don’t post “that’s crappy graphics!”. The thing that matters is even with 5 minute graphics the game is awesome. Click on video to watch in Higher Quality.
Made bots up until the point where I can play with them. I’m not happy with it, it’s not fun. The movement feels bad (the one from the Test Box I posted here). I slowed down the movement to make it more realistic. It is much better. Still something is missing.
I installed a profiler that works with VS 2003 C++. I never used anything like this before. I think this is the next best thing after a debugger. It’s brilliant I learnt a lot just by looking at the performance of different functions. I’ll be optimizing the code for the next few days.
We are discussing with Sigvatr some possibly brilliant new ideas. Remember my post about Seamless Servers?
More bot programming. I made them use the waypoints even if they are not walking on a path. That means if they see an enemy and between them there is a hole, they do not blindly go like a zombie at the enemy and fall into the hole. Instead they use waypoints to determine what is usually done at the spot they are on. So if they want to walk to an enemy but on the given spot players usually jump, they will jump. This is a very simple, yet very satisfying trick.
I made a bot today with working waypoints. Waypoints are recorded by the player moving then the bot uses that as a path for his movements. It looks promising. Maps in their architecture are more simple than in Soldat, so I think this will be sufficient for realistic bot movements. Although I am thinking of making some kind of ‘probability-based’ algorithm for the movement. It would be more life-like since things in nature work more like that. We’ll see, I’ll keep you updated.