How do you expect a game to respond?

(Andrew Plotkin) #21

I can’t tell from an animated gif how Jinxster deals with the design problems we’ve been talking about.

Does it take as a principle that the player shouldn’t be able to interrogate the UI (autocomplete, etc) about objects that haven’t yet been seen in the game? How does it deal with words that are adjectives in some contexts and nouns in other contexts?



The results of a very small amount of testing:

No. It is easy to probe the game’s vocabulary.



I didn’t read the entire thread so I may be missing some important aspects of the discussion. I just wanted to add the behavior I put in my changes to the Hugo library:

(which I guess is a simplified version of one of the original suggested possibilities)

Anyhow, maybe there are certain cases where this will sound awkward, but I felt it was a pretty decent compromise that doesn’t insult player knowledge.


(Andrew Plotkin) #24

Does it say the same if the box starts closed, so you’ve never seen the hat?

What if the hat is secretly teleported away after you close the box? (Yes, this is an intentionally silly edge case.)


(Andrew Plotkin) #25

(But it gets at the interesting question: how much player knowledge do you want to model for this sort of response?)



No, the hat must be known for that response (and in the same room, of course); I figure there might be cases where authors will want to teleport away objects to dictate whatever response they want to give.


(jkj yuio) #27

Hi Zarf,

The UI auto-complete is totally different from the parser and is not really related to the original issue.

But out of interest, the UI auto-complete is totally game independent. It reads the text and only ever suggests something that had been mentioned. That way it never gives anything away. Actually, it can’t anyway since it has no game knowledge.

Regarding adjectives and nouns, this is a game parser problem;

From my knowledge of Inform (i could be wrong), there is quite a woolliness here, around adjectives vs nouns. Jinxter (and Magnetic Scrolls systems in general), work by resolving words into either adjective and/or noun and/or verb. You can have a “red car” and a “red red”, supposing there was such a THING as a “red” as well as an adjective. thus;

“plant the pot plant in the plant pot with the trowel.”

You definitely have to resolve whether words are adjectives or nouns or verbs (above “plant” is all three).

Here’s an annotated transcript;

look yellow
What yellow? // yellow is a pure adjective.
look yellow door
What yellow door? // when there is none
look automatic door.
You can’t see an automatic door here. // something elsewhere.
look front door
The front door is a decent, solid-looking job of fumed oak, the householder’s joy and burglar’s despair. It is locked.

OK so the first thing here is that we can secretly find out if there’s a yellow door in the game. However, i don’t see this as a major drawback since you have a similar problem with unknown words;

Look cheese
You can’t see any cheese here.
Look helicopter
I don’t understand helicopter.