Slide 1
Slide 1 text
Operationalizing
Creative Theories
Chris Martens
UC Santa Cruz
Expressive Intelligence Studio
OBT 2016
1
Slide 2
Slide 2 text
2
Slide 3
Slide 3 text
3
Slide 4
Slide 4 text
4
Slide 5
Slide 5 text
5
Confluence
Slide 6
Slide 6 text
confluence of
programming abstractions
and
interactive expression
6
Slide 7
Slide 7 text
Roadmap
!
Imaginative Play
Collaborative Play
Generative Methods
Tool Design
7
Slide 8
Slide 8 text
Roadmap
!
Imaginative Play
Collaborative Play
Generative Methods
Tool Design
8
Slide 9
Slide 9 text
“Imaginative play: [children] must practice pretending
to be someone else, in another place. Making up
character, plot, and dialogue, for the stories they
invent. They are, in truth, inventing abstract thinking.”
Vivian Gussin Paley, early education
researcher:
9
Slide 10
Slide 10 text
10
Slide 11
Slide 11 text
Hypothetical Reasoning
11
Slide 12
Slide 12 text
Hypothetical Reasoning
wizards are real,
wizards can cast spells,
“accio” is a spell that summons an object,
…
!
12
Slide 13
Slide 13 text
magic circles
Adams & Rollings 2009
13
Slide 14
Slide 14 text
other “magic circles:”
!
metacircular interpreters
self-similarity & fractals
recursion in general
14
Slide 15
Slide 15 text
“magic” in digital games
15
Slide 16
Slide 16 text
16
inventory systems
Slide 17
Slide 17 text
17
verb/noun systems
Slide 18
Slide 18 text
emergent behavior and surprise
!
a response for every conjuring of the
imagination
!
game worlds in conversation with their
players
18
Slide 19
Slide 19 text
19
Slide 20
Slide 20 text
20
Emergent Behavior
Slide 21
Slide 21 text
21
Simulationist Worlds
e.g.:
glass can break into shards
rope is cuttable
shards are sharp
=>
glass can be broken to cut the rope
Slide 22
Slide 22 text
22
Simulationist Worlds
Slide 23
Slide 23 text
23
Inform 7
Slide 24
Slide 24 text
Roadmap
!
Imaginative Play
Collaborative Play
Generative Methods
Tool Design
24
Slide 25
Slide 25 text
25
Tabletop RPGs
Slide 26
Slide 26 text
Surrealist games
26
Slide 27
Slide 27 text
27
Like programming languages, collaborative
games are constrained affordances whose
use results in a designed artifact.
Slide 28
Slide 28 text
Chooser: decide among available rules
Interpreter: what does the output mean?
Interpreter: what do the rules mean?
Co-creator: contribute rules
28
The Role of the Human
Slide 29
Slide 29 text
Nomic
[Peter Suber, 1982]
29
Game with self-modifying rules
Slide 30
Slide 30 text
30
Blurring the lines between
design tool and designed artifact
Slide 31
Slide 31 text
Roadmap
!
Imaginative Play
Collaborative Play
Generative Methods
Tool Design
31
Slide 32
Slide 32 text
32
AKA Procedural Content
Generation (PCG)
Slide 33
Slide 33 text
Spatial level structure
Character attributes
NPC Dialogue
Items
Aesthetic components (visuals, music)
Narrative*
Game Rules*
33
What kinds of things can we generate?
Slide 34
Slide 34 text
Level Structure: Spelunky
34
Slide 35
Slide 35 text
Aesthetics: Secret Habitat
35
Slide 36
Slide 36 text
Dialogue: Prom Week
36
Slide 37
Slide 37 text
DESIGN FOR RECOMBINATION
!
[Gillian Smith, “An Analogue History of PCG”]
37
Slide 38
Slide 38 text
Oulipo
38
Slide 39
Slide 39 text
39
Slide 40
Slide 40 text
40
Slide 41
Slide 41 text
Lego
41
Slide 42
Slide 42 text
Compositionality
42
Slide 43
Slide 43 text
Combining Hypotheticals
43
Slide 44
Slide 44 text
Compositionality
44
Meaning Combination function
Slide 45
Slide 45 text
Compositionality
45
With [[-]] typing, value, etc.
Slide 46
Slide 46 text
production grammars
& rewrite systems
46
Slide 47
Slide 47 text
production grammars
& rewrite systems
47
Slide 48
Slide 48 text
Context Freeness
is a form of compositionality
48
Slide 49
Slide 49 text
49
- Given a combined thing, its meaning can be
determined by its parts
!
- Given an atomic thing, its meaning can be
understood independently of what it can be
combined with
Two Views of Compositionality
Slide 50
Slide 50 text
for any new domain we want to generate, the question
becomes:
!
what are the Lego pieces?
how do they compose?
50
Slide 51
Slide 51 text
Narrative
51
Slide 52
Slide 52 text
Vladimir Propp: Morhpology of the Folktale
52
Slide 53
Slide 53 text
53
Are stories compositional?
+ = ?
Slide 54
Slide 54 text
54
Slide 55
Slide 55 text
55
“18 Cadence” by Aaron Reed
Slide 56
Slide 56 text
56
A refinement of compositionality
If A’s interface I(A) matches I(B), then
I(A * B) is defined.
e.g.: function composition
Slide 57
Slide 57 text
57
E.g.
A hates B
A has
weapon
A murders B
B dead
B hates A
A insults B
B hates A A hates B
Slide 58
Slide 58 text
58
E.g.
A hates B
A has
weapon
A murders B
B dead
B hates A
A insults B
B hates A A hates B
Slide 59
Slide 59 text
B hates A
59
A hates B
A has
weapon
B dead
B hates A A hates B
A murders B
A insults B
Slide 60
Slide 60 text
60
symmetric monoidal categories
Slide 61
Slide 61 text
61
Slide 62
Slide 62 text
62
analogies for PCG
Slide 63
Slide 63 text
63
Can an automated system come up with coherent recipes?
Or coherent programs in a given PL?
(program synthesis)
Slide 64
Slide 64 text
Mixed-initiative design: Sentient Sketchbook
64
Slide 65
Slide 65 text
Roadmap
!
Imaginative Play
Collaborative Play
Generative Methods
Tool Design
65
Slide 66
Slide 66 text
Tools that lots of people
use are more fun when
they’re compositional
66
Slide 67
Slide 67 text
Twine
67
Slide 68
Slide 68 text
Tools that exist: Tracery
68
http://www.crystalcodepalace.com/traceryTut.html
Slide 69
Slide 69 text
Tools that exist: Tracery
69
Slide 70
Slide 70 text
70
6 6¶
6R6¶
U 6R6¶
Q
/RJLF3URJUDP
LQLWLDOVWDWH
Ceptre
Slide 71
Slide 71 text
…and compositional tools
are more fun when other
people can use them
71
Slide 72
Slide 72 text
programming a blank canvas
72
Slide 73
Slide 73 text
compiler
73
structure Foo =
struct
!
fun fact 0 = 1
| fact n =
n * fact (n-1)
!
end
REPL
Slide 74
Slide 74 text
compiler
74
REPL
fun fact 0 = 1
| fact n =
n * fact (n-1)
fact : nat -> nat
fun mapfact l =
map fact l
Slide 75
Slide 75 text
Agda emacs mode;
interactive theorem provers
75
mixed-initiative compositional
programming?
Slide 76
Slide 76 text
76
programming in conversation
with a system
Slide 77
Slide 77 text
CODA
77
Slide 78
Slide 78 text
<3 OBT <3
78
Slide 79
Slide 79 text
<3 OBT <3
2012
79
Slide 80
Slide 80 text
<3 OBT <3
2012
2014
80
Slide 81
Slide 81 text
Different Games : GDC
::
OBT : POPL
81
Slide 82
Slide 82 text
DIFFERENT GAMES are the
FUTURE OF GAMES
- Clara Fernandez-Vara,
Different Games 2013
82
Slide 83
Slide 83 text
OBT PL RESEARCH is the
FUTURE OF PL RESEARCH
83
Slide 84
Slide 84 text
Chris Martens
[email protected]
@chrisamaphone
http://www.cs.cmu.edu/~cmartens
84
THANKS!
Slide 85
Slide 85 text
Extra Slides
85
Slide 86
Slide 86 text
86
top-down approaches:
evolutionary, constraint, planning
(methods of search)
Slide 87
Slide 87 text
87
PCG with ASP
Slide 88
Slide 88 text
Simulationist story generation:
88
Describe what characters are in the world & their
relationships to one another
!
Describe what story events can do to change those
relationships
!
Describe a termination condition
!
Press play
Slide 89
Slide 89 text
storyteller
89
Slide 90
Slide 90 text
vi hart on structure:
!
"we understand the world and each other through the
patterns we know, but it's also important to realize
the structure is there, realize how invented it all is.
!
“that frees you to build new, more beautiful structures
that communicate things there were no words for
before, even if that new word might not be
understood."
90
Slide 91
Slide 91 text
magic circles?
Pfenning & Davies 2000
91