While most of ChatGPT-like AI models have issues with questions related to Inform due to hallucinating things that aren’t possible or are done in a different way, I’ve found that Perplexity is uncannily decent at producing Inform 7 code that works as well as answering general questions of How to do X in Inform 7 variety.
Which is actually a good thing since I won’t have to bother you with my noob level questions anymore.
I definitely wouldn’t worry about that – most folks here like answering newcomers’ questions, since many of us were noobs ourselves not long ago! And it’s also broadly helpful for the forum to be a repository for questions and issues folks run into when making their games; I know I’ve found the archives here a really valuable resource when I’ve run into something that doesn’t make sense – and if you have a question, I’m sure a bunch of other people have had, or will have, it too.
More broadly, while I haven’t seen the level of feedback Perplexity in specific is able to provide, Inform has a lot of nuances around things like the action processing sequence, or when to use lists vs. tables, that experienced users here can probably help spot and manage. So yeah, I’d definitely say don’t worry about bothering folks with questions that seem basic – they might actually be more complex than they seem, and this place is designed for those kinds of conversations regardless!
Just want to take a moment and really stress this point.
This forum exists to answer noob questions, so that other new users can find them and learn from the answers and discussion.
This is not a forum reserved for experts.
You are not causing a problem or annoying people by asking questions that you assume to have low-hanging answers.
Noob questions turn into answers that future readers can find and make use of.
I use AI to improve and correct my translations of texts into English, as English is not my native language. Regarding Inform 7 code, it is true that AI can provide some hints on syntax and point out an error that we can no longer identify correctly after a four-hour programming session, but its usefulness quickly becomes limited, in my humble opinion.
Indeed, since it’s about programming a game, especially an interactive fiction game, the mechanics of game design and narrative design are closely intertwined, and these two components are (or should be) closely linked to the underlying worldbuilding. The code is a synthesis of all these aspects.
For this reason, it is always beneficial to seek advice from seasoned veterans, because behind questions that may seem purely technical often lies the need for stepping back to understand the bigger picture.
For example, someone asking a question like, “How do I program a before rule to respond to a player who wants to examine a book that doesn’t exist in the library?” to an AI would get their before rule line of code. But an experienced human would advise them to use a check rule instead, because since the player has decided to examine the book, the examination process should be honored, even if it ends abruptly. In a complex game, this avoids the systematic evaluation of the before rule and subscribes to the action processing of examining books, which allows for more scalability in the long run.
That’s why a recurring question that veterans of this forum (of which I am not yet a part) often ask is a variation of this one: beyond your technical question, what kind of experience do you want the player to have?