week 11

For this week I will be learning some basic animation from one of the animators to understand the workflow of other fields and in doing so I have modeled some basic objects such as a table, chair and a few other things visible in the picture below.

Another learning outcome I have been working on is around making a API wherein I decided to make a dll file , the reason I chose a dll file was because this is something new to learn and it is a good way to protect my code from being used by unauthorized users .

To start off on learning how to make a dll, I started with looking at lots of guides but noticed a lot of them were not using monobehaviour which is something I needed to run my steering behaviors as they use man classes found in monobehaviour such as transforms and vector3.

After some assistance from my lecturer I have been able to find the right resources and start working on the dll itself but more things are yet to go wrong as a life of a developer constantly revolved around the concept of one step forward and two steps back.

I started working on my dll file in visual studio 2017 community edition with the target being the highest version of dotnet which is 4.7 , the latest version at the time of writing this post . After moving my existing code , I made the changes needed to my steering behaviors , this included decoupling the code from the existing parameters specified in the script by adding arguments to each behavioral state instead . A step I skipped to mention was that I needed to add the unity dll to the references in order to gain access to the unity based classes such as vector 3, game objects etc.

For my spatial portioning paper I am almost done, all I need to do is add a real world example of a project where spatial portioning was used to improve performance along with some basic pseudo code explanation to finish the paper and be done with the submission.

A small bug I discovered in our game was with in regards to collisions where in the projectile the player fired would sometimes go through walls, this was something very simple to fix by setting the rigidbody Collison detection to continuous and interpolation to extrapolate. The problem was that the projectile was going very fast and it would go through objects in between frames, this was because the rigidbody settings were set to discrete which was not able to predict the movement of the projectile in-between multiple frames.

Another minor bug we had was with collisions where in the player object would go through walls when it rotated in a corner which can be seen in the picture below.

The reason we had this problem was because each child of the player object such as shields and our barrel had a rigidbody and thus was detecting collisions with each individual child rather than the object as a whole , to fix this we moved the rigidbody to the object that all the items were chided under which fixed the problem .

Now that we have fixed most bugs, we will be preparing to publish the game for windows & android which is out exotic platform for this semester but we will be doing more bug testing before we do proceed to make sure the game works as is intended.