The last time I programmed in c++ was in 2015 in csci 1113. To create this graphics project I had to first remember and get the hang of c++, and am still doing so. The code was a bit intimidating the first time I looked at it. I had to spend a good amount of time getting a sense of what the variables were and what the functions do. Prof Guy showed us how to import part of a ppm, and finishing the rest wasn't too hard.

I implemented translate, and scale first. Translate was pretty much already done for us. Initially scale was hard because I didn't realize I had g_clicked_x and g_clicked_y, which are key. I think I discovered them when I was doing the normalization calculations for figuring out which edge the user clicked and dragged. After I saw the clicked variables, scale was easy to get working by trial and error.

Rotate was the hardest, mostly because I haven't done transformations since high school. Editing updateVerticies() was ok. I remembered the basic x'=x*cos(t)-y*sin(t) and y'=x*sin(t)+y*cos(t). I had to re figure out how rotating about a point other than the origin worked, x'=x_0+(x-x_0)*cos(t)-(y-y_0)*sin(t) and y'=y_0+(x-x_0)*sin(t)+(y-y_0)*cos(t).

Figuring out how to "undo" a series of transformations was most challenging. Thanks for the piazza help. I scale first, stored in vars c & d. Then I rotate by using -g_angle and translate and store the result in x & y, which are normalized. I use arctan to figure out the dragged angle, which isn't perfect because, depending on quadrant, the angle needs to be adjusted by -π. My code works, but isn't as robust as it could be (line 257). I'm not sure how to make it robust.

I created the picture. It's supposed to be a person trapped in a box.

Click here for zip file. Mac OS. run.sh will compile my code. square is my exe.