Trying to sus out 2014 Inform responses guidelines

I’m trying to work out when a response is required by the standards of the New Inform.

“Firstly, extensions should always use Responses to reply to the player’s commands:”

I get that all the standard rulebooks reporting on player actions use them, etc. But I feel a little fuzzy about other kinds of Inform entitites.

For instance, activities. Should things ‘said’ during an activity be responses?

And what about internal source actions which print text?

eg In my source, if I make an internal action for myself called ‘To gild the lily: (some code)’ (so I can shorthand some code), it’s not a rule is it? And thus I couldn’t make a response. But it is a chunk of an extension which may print text. And it will do so at different times.


I think - FWIW - that if your “To gild the lily” rule is part of your extension, then any text that that rule prints out should be a response that a user may easily customise. If it uses more than one Say command, every time you Say something it should be a response, even if all those commands are generating a single sentence.

EDIT - I can’t believe it. I’ve been playing so many misspelled Twine games that I actually had to come back and correct two “you’re” back to “your”. Aaaaargh!

EDIT 2 - I would imagine (again, not 100% sure) that this is the same for activities.

It helps if you think why this rule is in place. It’s probably so that a user can easily customise the text printed by an extension. Therefore, it stands to reason that every text the extension spews out should be a Response, which is currently very, very easily customisable.

The rule of thumb is: if someone is making a game in some other language than English, would they need to translate this text? If yes, it has to be a response.

In this particular case you could do:

To gild the lily:
    follow the lily gilding rule.

This is the lily gilding rule:
    say "Lily gilded!" (A).

Letting authors name to-phrases and allow responses used inside them might be worth a feature request.

I think the reason that responses aren’t needed for phrases is that they can be easily overwritten just by defining the phrase again. But phrases can be long with multiple texts, so extending the responses system to them would make sense.

Being able to name to-phrases would also be helpful from the perspective of the story author being able to easily replace the whole phrase (regardless of whether it contains responses) without worrying about the order in which things appear in the source (discussed in this thread).

Even if this feature is not added, I think it’d be nice to have responses/extension guidelines spelled out in more detail somewhere, to answer questions like this:


And from this thread:

Where to put the (A) in relation to the string. For instance, this works

But this does not:

Other questions I have asked:

What about debug text that will appear for the author but not in a released game? Does that need to be lettered, and in named rules?

Does this sort of text need to be adaptive: “This game [offer] feature X, and [we] [can see] it by typing Y”.

I’m pretty sure that the rule is that out-of-world messages don’t need to be adaptive to story viewpoint and tense. See the discussion on this bug with Word of Graham in the comments. A message like “I only understood you as far as…” wouldn’t be adaptive, because it’s a message from the parser to the player. A message like “It is pitch dark, and you can’t see a thing” needs to be adaptive–in a first-person past narrative, it needs to be “It was pitch dark, and I couldn’t see a thing,” because it’s the player who can’t see a thing.

So, I’m pretty sure “This game offers feature X, and you can see it by typing Y” should not be adaptive, because it’s a message about the game rather than in it.

Thanks, that was my understanding as well. I did ask it at one time, and got it answered…my purpose for mentioning it here was mostly to put the questions-that-people-have-asked-or-been-confused-about in one place. It seems like enough questions come up that a FAQ or guide somewhere or more explanation/examples in the manual would be useful. (I’d be happy to assemble the answers I’ve found all in one place, if that would help anything.)

Thanks for gathering up the info, bg.


It’s definitely worth a feature request if the alternative is as horrible as what you suggested here :stuck_out_tongue:

I’ve got ‘to (blah)’ phrases which call further ‘to (BLEH)’ phrases, and if these all become further nested with redundant rules so that they can have responses, the code will quickly begin to disappear up its own fractal posterior.