Every year I update this graph of how many IFComp entries there are and how many are parser-based vs choice-based.
Why? Obviously this is a silly, pettifogging way to view IFComp. On the other hand graphs are nice. And if somebody says “parser IF is dying!” it’s fun to be able to point at a chart and say “Yes, the trend indicates that parser IF will be gone by the year 2028! Enjoy the next twelve years of parser games, you fossilizing relic!”
Which is what I would have said last year. But for the past few years, the parser count has been creeping up again. Even the smoothed (dotted) curve has stopped decreasing.
Now, I said this was a silly and reductive view. One way it is silly is that the definition of “parser” just isn’t that obvious any more. This year there are two games written using Texture, which does not have a parser but does have a noun-verb selection interface that supports a parser-like world model. The same goes for Detectiveland. Then there’s Fallen Leaves, which could be described as a verb-adverb selection interface. On the flip side, there’s Take (parser, but only accepts one verb) and Mirror and Queen (parser, but only accepts single words).
I decided to take a broad view and accept all of these as parser-like. For the splitters, the totals look like this:
3: reduced parser
5: parser-like (at least two axes of choice range)
So my graph shows 26 parser-like (blue) and 32 choice (green), but if you want to be strict you could call it 21 to 37. (I am confident that reduced-parser games have been counted as “still a parser game” throughout IFComp history.)
(I should note that I have not gone back to earlier years to apply this notion of parser-like-ness. Sorry. Some other time.)