When I was young and first started out I only wanted to use C because I thought C++ was cheating. I also wanted to create everything myself so I had 'complete control'. While the learning experience was good I got pretty much got nothing done as I was fiddling with tech the entire time. It was very hard for me to get over that feeling. It took getting my first real job to finally get over it. Spending two hours tweaking code or XML to create a a puddle when you could have used a slider in 30 seconds is a complete waste of time. Unless fiddling with tech is what interests you when programming. For me it doesn't, making a fun game that other people can enjoy is far more important to me then writing yet another specular lighting shader that is just like the other hundreds that are out there.

The whole vibe I get from this thread is looking for justification to do thing the 'hard way'. If you want to create all the puddles in your level with code then by all means do it. But do realize that if something changes in the level now you have to go back and manually fix all those puddles. If you took the 'easy way' with a visual editor you just click the puddle and move it. You're not making yourself any less of a programmer by doing this. You are saving yourself hours of menial labor. The 3D Breakout clone I made, the first couple of levels were done by hand in a text editor. It took hours to get a level setup as had to tweak a value and then run the game to see if it was anywhere close to where it should be. After a few levels of that I took a few days and created a level editor with C# and SlimDX. What took three and four hours now took 30 minutes and I could now let my artist make the levels as he was much better at that kind of thing. This is the exact kind of thing that happens in a game studio. When you are paying 30 guys $50k+/year you don't want them wasting time manually tweaking variables all day. Unless you are working for a middleware company then most likely you aren't creating game engine stuff from scratch. In the long run you are doing yourself a favor by getting used to do things the 'easy way' as thats what you will do at the job.

If making the low level engine stuff is what excites you then you should go with it, Java and Slick2D is a good start to that path. For all the hordes of people like me that would rather use some kind of game engine there has to be a dedicated few that know the ins and outs of the target hardware to be able to create that engine to run smoothly. The most important thing is to just pick something and go with it. Its not like the path you pick today has to be the same thing you are doing in 20 years.