>>32
In theory, you don't write any actual Python when scripting a Ren'Py game. All the dialogue, animations, and even UI layout are supposed to be done through a custom language that is parsed and expanded into a graph of Python objects. The engine traverses this graph to predict what assets need to be loaded at what points in the story, thus allowing you to play through the whole game without ever seeing a loading screen. This is not a bad idea on paper, and any competing VN engine should more or less aim for this same design.
The problems are with the implementation. The documentation is extensive, but not thorough, largely because the DSL layer is a leaky abstraction. You can "see through it" to numerous internal details of the engine itself. Interacting with some of these does involve writing Python. Understanding their behavior usually involves reading the source code. In some cases you have to write C and rebuild the entire engine to change how things work.
Introducing another language to this mess would not fix anything. Use Ren'Py with its DSL or throw it out completely and write something better from the ground up.
people are actually considering using python for this
>>38
I'm >>7,10,18,34,37 (and a few others).
There are always tradeoffs to be made between doing something correctly and just getting it done. As I have said, I am against it both on principle and because of past experience with its flaws, but if people are so eager to get this thing made that they can't wait another couple months to see some text superimposed on a background image, then fine, go ahead and use Ren'Py and make a proof-of-concept. You could even use PowerPoint for all I care. Port it to an engine written in Scheme once you run up against the limitations inherent in the tools that are close at hand.
If people can have slightly more patience, but can't contribute enough effort to write it in Scheme, maybe Common Lisp is a better compromise. That Shinmera joker made a game with SBCL. I've had my eye on the dialogue engine for a couple months. It wouldn't be a bad starting point.
>>39
I don't think anyone should reasonably assume the whole game to be written in Scheme, despite the major hack value that might bring; having parts of the game utilize an embedded Scheme interpreter would be novel enough, IMO.
If it turns out Ren'Py truly is too limited for our needs and if someone actually conjures up a usable alternative, then we should surely make it happen.
As for using CL I am much more familiar and comfortable with it than Scheme, so there's 1+ potential programmers.
In other news, I made an IRC channel for this board, as evidenced here: http://textboard.org/prog/715
Assuming there's real interest present we could house less pedantic discussion there if need be. I'll try to be as responsive as I can.