SLIDE 1
Immutability,
- r Putting the Dream Machine to Work
Immutability, or Putting the Dream Machine to Work The trie memory - - PowerPoint PPT Presentation
Immutability, or Putting the Dream Machine to Work The trie memory - - PowerPoint PPT Presentation
Immutability, or Putting the Dream Machine to Work The trie memory scheme is ine ffi cient for small memories, but it be- comes increasingly e ffi cient in using available storage space as memory size increases. The attractive features of the
SLIDE 2
SLIDE 3
SLIDE 4
SLIDE 5
The trie memory scheme is inefficient for small memories, but it be- comes increasingly efficient in using available storage space as memory size increases. The attractive features of the scheme are these: 1) The retrieval process is extremely simple. Given the argument, enter the standard ini- tial register with the first character, and pick up the address of the second. Then go to the second register, and pick up the address of the third, etc. 2) If two arguments have initial characters in common, they use the same storage space for those characters.
- J.C.R. Licklider, "Man-Computer Symbiosis" 1960
SLIDE 6
SLIDE 7
SLIDE 8
SLIDE 9
SLIDE 10
SLIDE 11
Model-View-Controller
SLIDE 12
๏ first formulated by Trygve Reenskaug
Adele Goldberg and others at Xerox PARC in 1979
๏ long shadow, the basic concepts still
prevalent today.
SLIDE 13
๏ At a very abstract level MVC is a sound
separation of concerns
๏ Implementations leave much to be
desired
๏ Stateful objects everywhere
SLIDE 14
SLIDE 15
SLIDE 16
Functional Programming and Data
- immutable values, not mutable objects
- “change” returns a new value, leaving the old one unmodified
- they’re persistent
- they’re fast
SLIDE 17
Simple example: Linked List
X
SLIDE 18
X Y
Simple example: Linked List
SLIDE 19
X Z Y
Simple example: Linked List
SLIDE 20
X Z Y
structural sharing
Simple example: Linked List
SLIDE 21
Sharing structure
- space effjciency
- computational effjciency – avoids copying
SLIDE 22
Phil Bagwell
- Array Mapped Trie
- Hash Array Mapped Trie
SLIDE 23
Bitmapped Vector Trie
- data lives in the leaves
- e.g. prefix tree used for string lookup
- bitwise trie
SLIDE 24
Persistent Vector
SLIDE 25
Persistent Vector
SLIDE 26
Persistent Vector
SLIDE 27
Persistent Vector
SLIDE 28
Persistent Vector
108 109 110 111 104 105 106 107 100 101 102 103 96 97 98 99
SLIDE 29
Persistent Vector
getindex
SLIDE 30
Persistent Vector
108 109 110 111 104 105 106 107 100 101 102 103 96 97 98 99
SLIDE 31
Persistent Vector
108 109 110 111 104 105 106 107 100 101 102 103 96 97 98 99
0b01101010
SLIDE 32
Persistent Vector
108 109 110 111 104 105 106 107 100 101 102 103 96 97 98 99
0b01101010
SLIDE 33
Persistent Vector
108 109 110 111 104 105 106 107 100 101 102 103 96 97 98 99
0b01101010
SLIDE 34
Persistent Vector
108 109 110 111 104 105 106 107 100 101 102 103 96 97 98 99
0b01101010
SLIDE 35
Persistent Vector
108 109 110 111 104 105 106 107 100 101 102 103 96 97 98 99
0b01101010
SLIDE 36
Persistent Vector
assoc
SLIDE 37
Persistent Vector
108 109 110 111 104 105 106 107 100 101 102 103 96 97 98 99
SLIDE 38
Persistent Vector
108 109 110 111 104 105 106 107 100 101 102 103 96 97 98 99
SLIDE 39
Persistent Vector
108 109 110 111 104 105 106 107 100 101 102 103 96 97 98 99
SLIDE 40
Persistent Vector
108 109 110 111 104 105 106 107 100 101 102 103 96 97 98 99
SLIDE 41
Persistent Vector
108 109 110 111 104 105 106 107 100 101 102 103 96 97 98 99 104 105 foo 107
SLIDE 42
Persistent Vector
Length 4 internal vectors?
SLIDE 43
Persistent Vector
32
From Bagwell, Rompf 2011
SLIDE 44
327
SLIDE 45
34,359,738,368 elements
SLIDE 46
demo
SLIDE 47
SLIDE 48
SLIDE 49
SLIDE 50
Om
SLIDE 51
SLIDE 52
f(D0) = V0
SLIDE 53
f(D1) = V1
SLIDE 54
diff(V0,V1) = CHANGES
SLIDE 55
SLIDE 56
demo
SLIDE 57
SLIDE 58
Persistent Data Structures … ROCK
SLIDE 59
SLIDE 60