Still nothing playable, but fair progress by my usual standards.
Last time, I was planning to work on the scripting for the animated sequences. However, when I sat down on holiday to have a good think, that instead became the starting point for a virtual machine. It doesn't do much at the moment, but already feels like the better way to make something that can be played, rather than watched.
As far as I can remember, in the process of bashing out a tiny Spectrum prototype for someone else, I realised just how difficult it was to follow numerical variable references in my current scripting language. While working out on paper that I could use real names instead, if I then substituted them for numbers at compilation time, my attention wandered back to making a little world that players could explore.
Once back at the computer, but still working partly on paper, I realised that the scripting language was also a poor fit for constructing a navigable map. On a whim, I discovered that not only was XML potentially much easier for people to read, it was also well supported by Python.
While writing a Python routine to convert my XML into Spectrum data, I realised that, having previously followed what seemed like a dead-end with PyGame, it wouldn't require much more work to visualise this data.
Then, by the time that was done, including keyboard commands to skip between screens, and a timer to sleep when not redrawing anything, it seemed silly not to try prototyping the game this way.
There's nothing playable yet. It's not even a true virtual machine so far, as it's working from native data structures, rather than unpacking bytes and bits as required. However, I can read the code back without simultaneously struggling to remember what it does, which is quite an improvement on my scripting language.
Email: comments at arbitraryfiles.com