I tried something similar to this when I was at University - the idea was to make a class full of University students who had any range of ability, turn them into a model game development team, and then function as producer / programming lead to create a game in one semester (I had a co-teacher who was my co-producer and the art lead). In the end, we had about 6 to 8 people programming, 10 artists, and then a few people who couldn’t do either so did things like writing story and item descriptions. The idea was to make an RPG, mainly because it is very easy for people to work individually on something like an RPG, and there are more jobs than just programmer / artist.
It turned out that of all the programmers literally only one had any notable programming experience whatsoever (he was a peer of mine), but he was also one of the laziest people you’ve ever seen and so did practically nothing. So, I was pretty much teaching people to program and also trying to get a game out at the same time, a very similar situation to what you’re looking at.
The medium of choice was to make the game in Second Life, because it includes 3D tools, networking, physics, avatars, etc. all out of the box, and the scripting language is relatively simple to understand. Students were also able to see results from their programs almost immediately, which was a plus. In hindsight, I wouldn’t use SL again, because it is a painfully shitty program - ultra laggy and we needed to pay fees of about $40 / month to maintain the virtual land which hosted our game.
In the end, however, we did have a game. It was definitely impressive. It had cutscenes, monsters, boss battles, several fully-created worlds, and more. It didn’t function very well (SecondLife running like crap was actually the main cause of this), but it was a massive accomplishment for all involved. I was plenty pleased.
Now comes my advice to you: in order to make the above happen, I had to be putting in 5 to 10 times as much work as the students, in and out of class. In the amount of time I spent helping the programmers along, teaching them, fixing bugs, etc., I could have made the same game and I could have done it much better (just the code, I mean). In fact, much of the game was effectively mine. I made a lot of sample code that the coders directly used. I came to help people and give them implementation ideas often enough that, over time, these small suggestions and changes added up to me doing almost the entire piece of code. But, the students felt that they were doing the work and that’s what I wanted - I had a wonderful time. And I was impressed several times with one or two of the particularly adept programmers, who would go and do something cool on their own. So the lesson is that the quality of whatever comes out of this class is going to directly correlate to how much you you put in. Cas hit the nail on the head when he told his story about his coworkers. A lot of people just don’t understand game design/programming - it really is a specialized field. A lot of your students will never understand it; they will need to be led along with a carrot on a stick the entire length of the course.
Oh yeah, I also spent the entire semester before (just as many hours), creating sample code, a game skeleton, lesson plans, a schedule with deadlines, and an overall mold for the project to take. As a teacher I’m guessing you do this sort of thing anyway, but this was also very very important to the relative success of the course.
So that’s that I think, there’s my story. In the end every single student, I think, had a very good time, even the ones who found it wasn’t just a slacker course. 
I’d say screw your principal and forget having a polished product, just do whatever you can to make sure your students feel accomplished and that they’ve learned something (and don’t use SecondLife!). In the end, if you keep it fun, the product will be impressive because students will be spending lots and lots of outside hours doing work.
Here are some videos of the end result of the class (it no longer exists because I am not about to pay $40 / mo to keep it around).
The Spacebar Time Continuum
Especially when you die and actually go to hell.