Although for the long run, my answer is yes, and I want to really experiment with Gen, I was wondering how practical it would be to try and learn and implement Gen into a project due mid-May.
The main reason behind wanting to learn and implement it into my current project is it’s 3D modelling potential.
I have recently been experimenting with morphing jit.gl.gridshapes, and I’ve been trying to do more than modulating them with sine waves, movies, and each other, as none of these methods yield enough control over the final shape.
I’ve looked into using jit.expr, jit.bfg to modify jit.gl.isosurf, jit.gl.nurbs, and jit.gl.mesh, and these looked like much better routes to modify 3D geometry, but I need to investigate and explore them more.
However, through my exploration into Gen, I saw this tutorial (Day 5);
There are a lot of things that Gen makes possible that otherwise wouldn’t be, but it also makes a lot of things easy that would otherwise be hard. Doing geometry mesh matrix manipulation is totally possible with jit.expr,bfg,op,etc., but the process of getting there is going to involve more conceptually difficult techniques. My personal experience with Gen is that I try a lot more stuff because it’s faster and often less headache than managing the equivalent collection of jit.expr objects. Also, with codebox in Gen, you can often directly port algorithms you find elsewhere for mesh deformations with just a few changes.
I knew that Gen would be more efficient in terms of real-time performance, however I didn’t think about the tidiness of the patch, which as you say can cause headaches and restrict experimentation and development.