Yes, source is the best we have.
It occurs to me that Craverly Heights is perhaps an even better candidate. This page with an annotated side by side comparison of its Inform 7 and Dialog source makes the intended behavior fairly easy to follow.
Discussions like this one crop up on a recurring basis…
- Looking for a small, open-source parser game
- Medium sized open source text adventure code?
- Simple story idea to test my engine?
- Seeking demo frameworks for inclusion with a FOSS IF Engine
- Cloak of Darkness 2
- Cloak of Darkness 2: Survey Results
- Standard Demo Kit for modern IF systems?
- Compare my new engine to others via ports
- Cloak of Darkness, Ring of Darkness - #9 by 8bitAG
But it keeps not happening. I suspect it’s because the domain is just too big. The spec calls for the moon to be visible in the outdoor locations. But does it dictate that the game should report an error if you try to give moon to aunt may? Does it specify what the error message should be?
In general, parser game devs rely heavily on existing standard libraries and its default world model, and errors, and messages. And no two of them have exactly the same world model. A specification should be, well, specific enough that one could document a test suite that determines whether something fails to implement the spec or (possibly) succeeds at implementing the spec. But there are so many different kinds of failure in parser IF that it would be extraordinarily tedious to create the spec, or document the test suite, or attempt to satisfy such a spec. I suspect that these are the rocks that aspirations to create such a thing crash upon.