In quite a few systems, the vision of a character is determined from the edges and corners of a character's square, not just the center. For instance, here is the dynamic lighting layer of a map. With this setup, here is what the character sees right now. Here is what the character should be able to see in many systems. Ideally, it would be best to have vision extend from everywhere on the token. However, I'm pretty sure that the dynamic lighting works by drawing individual vectors, and seeing from every part of the token would require a fundamental change in the way that dynamic lighting is set up. I could be wrong about that. As it is, this will requite four times as much work to draw the lighting from four different points (or five points if the center is still included), which could slow things down quite a bit if the lighting setup is complicated. If the option to Only Update on Drop isn't checked, then having the lighting update as a token is moved around would also be noticeably slower. There is currently a workaround for this. Hold down ALT and move your token to the corner and edges of your square. However, this is tedious. It might be necessary every time your token moves. Some GMs do not understand why this is necessary, and think that players are trying to cheat, meaning that a conflict between the platform and the rules is causing friction in groups. It's also possible to move your token a bit too far, and thus gain sight of something you aren't supposed to. Many people represent flashlights on tokens with a cone of light, and I admittedly don't know how this would work. It may be necessary for this to come from the center of the token only. Having the ability to shine a flashlight around corners would be adding another level of complexity to this. You may have picked up on the fact that I'm proposing that vision come from the edges of a character's token, rather than the corners. This because it could potentially prove to be problematic when tokens rotate, if one of the sources of vision is suddenly on the other side of a wall. Below, you can see that the bottom-right corner is on the other side of the line when the token is rotated clockwise. This could possibly by addressed by cutting off vision to a corner if there is any obstacle between the center and the corner, though I'm not sure how feasible this would be. (These next two images are .gifs, and you can click on them to view the animation.) Even if vision were done from the edges of a token, the issue of rotation would still need to be addressed. Someone could rotate their token, move it close to a border, and then rotate it back upright, as shown below. You can see that all of the squares on the edge of the token are above the line when tilted, but when it is rotated upright, the bottom edge is below the line. The reason I propose using the edges for vision is so that if the fix to this problem comes through preventing rotation of tokens, every token will still be able to rotate in place. I also speculate that there could be some rather weird bugs if the vision is generated from the corners rather than the edges (or possibly one pixel inside the edge, which would make little difference at the edges, but a significant difference on the corners in terms of angles of vision being blocked off). In light (heh) of the increased requirements and the fact that not all systems may be like this, I'm proposing a checkbox option rather than a site-wide change.
