Very interesting stuff, thanks a lot :)
The "detecting rotation or flip" for rules is pretty brilliant, not sure yet if "good brilliant" or "bad brilliant" but time will tell ^^
I'm wondering if it would be possible to adopt the same logic for the object definition?
Right now (talking about the ==== OBJECTS ==== section only): if I understood well, it's only "full package with rotation" and it makes the autocomplete menu quite hard to read and the autocomplete maybe a bit too much for some cases :
I feel the more frequent use cases are :
1-
URDL are 4 objects with a 90° rotation (like you did)
and
2-
Left / Right are mirrored, while Up / Down are either the same item or two different items (back and front, respectively)
exemple for same Up / down object : https://www.puzzlescript.net/play.html?p=25ee2f1c9000620aa7b6f429c958d095
exemple for different Up / down object (back / front): https://www.puzzlescript.net/play.html?p=2fad7fdd93a1a9d6f689024baeb84f5b
So you covered case 1- nicely, but didn't cover case 2- (or did I miss it?)
In case 2- what would be useful is not a rotation but rather a left/right asset flip (mirror)
(I believe up/down being mirrored versions from one another is much rarer?)
I feel that if you apply the same logic than you did in rules, i.e. defining one object at a time, it would make the thing both more readable and often more convenient.
Like if I define player_left then player_right, you could suggest a mirrored version which is probably 99% what people would want.
I reckon if I define Player_Up then Player_Down, I'm not sure what I'd want, but I guess mirrored version is the best bet too.
Now if I define player_up then player_right, then I guess suggesting all 4 would work. However, autocompleting only one object (player_right) might be more convenient (for instance I can define player_up, then player_right (autocomplete rotation) then player_left (autocomplete mirror) then player_down (autocomplete_rotation?))
it might also feel more consistent with what you did with rules (one at a time, the previous instance defining rotation / mirroring)
Sorry for not thinking that throught when you first asked, it came when testing :-) (also you delivered much to quickly :-D)
Anyway that's just my opinion from rapid testing, i'm actually quite excited to use the autocomplete in action, and I'm sure you've thought about all this much longer than I have so I'll trust your judgment.
Cheers,
Sévan.
Very interesting stuff, thanks a lot :)
The "detecting rotation or flip" for rules is pretty brilliant, not sure yet if "good brilliant" or "bad brilliant" but time will tell ^^
I'm wondering if it would be possible to adopt the same logic for the object definition?
Right now (talking about the ==== OBJECTS ==== section only): if I understood well, it's only "full package with rotation" and it makes the autocomplete menu quite hard to read and the autocomplete maybe a bit too much for some cases :
I feel the more frequent use cases are :
1-
URDL are 4 objects with a 90° rotation (like you did)
and
2-
Left / Right are mirrored, while Up / Down are either the same item or two different items (back and front, respectively)
exemple for same Up / down object : https://www.puzzlescript.net/play.html?p=25ee2f1c9000620aa7b6f429c958d095
exemple for different Up / down object (back / front): https://www.puzzlescript.net/play.html?p=2fad7fdd93a1a9d6f689024baeb84f5b
So you covered case 1- nicely, but didn't cover case 2- (or did I miss it?)
In case 2- what would be useful is not a rotation but rather a left/right asset flip (mirror)
(I believe up/down being mirrored versions from one another is much rarer?)
I feel that if you apply the same logic than you did in rules, i.e. defining one object at a time, it would make the thing both more readable and often more convenient.
Like if I define player_left then player_right, you could suggest a mirrored version which is probably 99% what people would want.
I reckon if I define Player_Up then Player_Down, I'm not sure what I'd want, but I guess mirrored version is the best bet too.
Now if I define player_up then player_right, then I guess suggesting all 4 would work. However, autocompleting only one object (player_right) might be more convenient (for instance I can define player_up, then player_right (autocomplete rotation) then player_left (autocomplete mirror) then player_down (autocomplete_rotation?))
it might also feel more consistent with what you did with rules (one at a time, the previous instance defining rotation / mirroring)
Sorry for not thinking that throught when you first asked, it came when testing :-) (also you delivered much to quickly :-D)
Anyway that's just my opinion from rapid testing, i'm actually quite excited to use the autocomplete in action, and I'm sure you've thought about all this much longer than I have so I'll trust your judgment.
Cheers,
Sévan.