Hints and quick post mortemPost-mortem (did I say quick? I lied)
Another pyweek, another set of gray hair. Why do this to myself? If I knew that I would be a much wiser man. Our plan of making an adventure game this time around backfired on many levels, but we still managed to complete something that I am fairly proud of.
The biggest problem we had, it turned out, was that while I was happy the first few days that coding the engine was going so well, I neglected to plan for all of the content and logic that we would have to put in to make the game actually work later on. Not that functional programming is bad, but I lapsed too much into funcionalism for different kinds of logic in the game that ended up making creating the puzzles and inserting objects into the screens very time consuming. Next time: focus on tools to help make adding content very easy. Last time we had a decent level editor that helped a lot, and though it was a very different type of game, I could have gone further to make scripting it more fun.
The second problem was that we chose an art-heavy game, and didn't have the time to work on it to really make a good attempt with the art. Some things look good, other things look bad, and none of the art is very cohesive. Still, I think continuing to choose art-heavy games is good in the long run, since it is the area we most need improvement.
Lastly, mercurial. My teammate and I, for the most part, were both working on the same code file. This would have been extremely painful in svn, but even with hg it probably would have been better to separate our work. It turns out tortoiseHG is just far too slow for rapid development - having to rightclick all the time and go to specialized menus for things. By the end, we were mostly using the command line, and it was a bit better. Having to commit, pull, merge, commit, and update on both sides whenever there were conflicts though, was really annoying. An automatic option to "synchronize" with a remote host, and make both sides the same, would be very welcome. Obviously, in other situations, the ability to do different operations independently is useful, but in this case, not so much. Before next pyweek I am going to look at writing a script to automate this process.
Most of the puzzles in Yard Sale are straightforward, but a few of the chains were short-circuited due to lack of time, so a few things are very senseless. The best advice is to make sure you have looked at and interacted with everything, and used everything on everything. If you follow that advice, you should have no problems completing the short "story". Here are a few more specific hints:
Don't forget to pick up the briefcase on the sidewalk, and remember the hobo.
I might post a walkthrough later if people ask.