I would always first look at the sources, then decide whether or not to base a new project on them.
Also I see Deu2000 present a project he wants to do about reimplementing the AGS engine runtime to Java. That's good, but it's nonsense he didn't look at the AGS sources to understand how the runtime works first.
Maybe you should analyze how AGS works first, then document it and finally reimplement it? You can go the fuzzie's way too (looking at the original messy code, then reimplement it), it's how extreme geeks do the things.
But you can't reuse any part of the original code at all, so it's like a complete rewrite.
Do you have the skills to start this project by yourself? Are you able to understand the AGS engine and document the runtime's behaviour? What do you think are the main benefits of this?
Maybe you aren't sure about the project benefits yourself, or you want others to follow you on it. Unfortunately that doesn't work, you must believe in your own project and start the challenge yourself first while asking others to help you.
PS: I'm from Spain too. And despite my English needs lots of improvements, it's better than yours. You should improve it if you want to collabore with other people