It was a cold and rainy day today so I decided to rework the parser of my text adventure game skeleton.
So now you can enter a simple sentence as can be seen in the screenshots which is better than the two-word approach I initially implemented (still not Infocom like but then that's not what I am after).
What happens in the background is that the user input is split into sub-strings which are then checked against a list of strings (e.g. object names, verbs), leaving their corresponding list index number on the stack. These numbers are then used within the actual game logic.
Oh, and I've added some colors - now the thing is beginning to look like a game ...





