I’m writing this mere minutes after my previous post, because while the topics are related, they focus on different aspects of the same thing: how bullshit preconceived notions on how we assume things work really are.

There was a great article a few years ago about how Fallout 3 handled a train ride. As players, we see a train in-game and I think our natural inclination is to believe that what we see is what we’re getting: a train, on a track (or at least on a ground plane) that follows a path set down by developers and designers. We see the ground and the train, and expect that it behaves for the most part just like a real-world train would. In Fallout 3, however, things weren’t always what they seemed.

No player would ever think that this is how the design was shipped, but there we have it. We look at this and we laugh and can’t even understand what the hell the devs and designers were thinking because it’s just too damn weird. Why not just make a working train? Although we probably understand that what was done was done for a reason, we can’t fathom that reason without explanation.

Bottom line: creation is messy. This “train head” (and the related “train motion hat”) works as a technical solution. The artistry, then, is making things look like we expect them to despite the mechanics not being even close to what we’d ever dream of. Had we never learned of this weird Fallout 3 situation, we all would have gone to our graves believing that we understood that the train in that game was, in fact, a train as we understand it.

In watching all of the Blender videos that I have watched to date, I was always under the impression that the way to hard-surface model was to start with a 3D object, to push and pull vertices, edges, and faces, and to work from that basic geometry. I always wondered how super-complex models could be sourced from a single cube or a cylinder because I couldn’t follow the path from one to the other. This workflow is how 98% of all videos I’ve seen (and I have seen a lot) start out, and how they progress. They describe all of the tools and techniques in terms of existing solid geometry so if I have an idea in mind for something to model, it’s real easy for me to believe that starting with that lowly cube is the best — and only — way to go.

Now, however, I have seen videos where the instructor traces concept art using only vertices which is then morphed into faces and ends with edges extruded in 2D. From a novice perspective, this seems like it violates the rules of 3D modeling somehow: it’s 2D tracing, even though the final shapes are extruded and result in a 3D model. Those complex models I mentioned earlier, I have learned, aren’t just one mesh, but maybe dozens of meshes created somewhere in the scene and brought together like a virtual LEGO set so that the final product looks the way we expect it. I suppose this is a “no duh” epiphany for some (or for some who want to look smarter than they are, because Internet), but for me it not only made more sense, but also cleared up a whole lot of potential stumbling blocks because I no longer have to think about taking what’s in front of me and pushing and pulling it to get the desired results. I can buck my own sensibilities and do what I feel necessary in order to achieve my vision (so long as I clean up after myself later on).

This, I think, is where a lot of instruction on the Internet falls down. There’s a lot to be said for giving people the tools to get their work off the ground, and then encouraging them to fly free with their knowledge, but my problem is that I tend to only take away only as much as I am given. At some point there’ll be a situation not covered by the basic overview of the tools provided in specific circumstances, and I feel that I am left with only the limited options covered by the videos I’ve watched. It doesn’t help that there are a lot of videos that only cover the “getting started” phase and rarely move into actual techniques. Playing around with things is certainly an option for further self-learning, but it takes time. Finding advanced instruction is also an option, but is hit or miss, difficult to find the specific info needed to fill the gaps, and can be expensive. Questions can be asked on forums, but that’s also hit or miss (will a kind, knowledgeable person actually answer, or just someone trolling with “maybe you should RTFM…”).

Understanding that the end result does not depend on a straight and logical path from start to finish is something I’m coming to grips with. Looking at a finished model and assuming that everything started from a single germ, or that 3D modeling is always 3D modeling weren’t even tools I knew I could leverage, but once I understood that nothing is off the table, so many options are becoming possible.