aboutsummaryrefslogtreecommitdiffhomepage
path: root/plugins/extraction/README
blob: 458ba0deb7571348a06d70e0218fa80f3a999013 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147

          Coq Extraction
          ==============


What is it ?
------------

The extraction is a mechanism that produces functional code
(Ocaml/Haskell/Scheme) out of any Coq terms (either programs or
proofs).

Who did it ?
------------

The current implementation (from version 7.0 up to now) has been done
by P. Letouzey during his PhD, helped by J.C. Filliâtre and supervised
by C. Paulin. 

An earlier implementation (versions 6.x) was due to B. Werner and
C. Paulin.


Where can we find more information ?
------------------------------------

- Coq Reference Manual includes a full chapter about extraction
- P. Letouzey's PhD thesis [3] forms a complete document about
  both theory and implementation and test-cases of Coq-extraction
- A more recent article [4] proposes a short overview of extraction
- earlier documents [1] [2] may also be useful.


Why a complete re-implementation ?
----------------------------------

Extraction code has been completely rewritten since version V6.3. 

1) Principles 

The main goal of the new extraction is to handle any Coq term, even 
those upon sort Type, and to produce code that always compiles. 
Thus it will never answer something like "Not an ML type", but rather
a dummy term like the ML unit. 

Translation between Coq and ML is based upon the following principles: 

- Terms of sort Prop don't have any computational meaning, so they are
merged into one ML term "__". This part is done according to P. Letouzey's 
works [1] and [2].

This dummy constant "__" used to be implemented by the unit (), but 
we recently found that this constant might be applied in some cases. 
So "__" is now in Ocaml a fixpoint that forgets its arguments: 

           let __ = let rec f _ = Obj.repr f in Obj.repr f


- Terms that are type schemes (i.e. something of type ( : )( : )...s with 
s a sort ) don't have any ML counterpart at the term level, since they 
are types transformers. In fact they do not have any computational
meaning either. So we also merge them into that dummy term "__".

- A Coq term gives a ML term or a ML type depending of its type: 
type schemes will (try to) give ML types, and all other terms give ML terms.

And the rest of the translation is (almost) straightforward: an inductive
gives an inductive, etc...

This gives ML code that have no special reason to typecheck, due 
to the incompatibilities between Coq and ML typing systems. In fact
most of the time everything goes right. 

We now verify during extraction that the produced code is typecheckable, 
and if it is not we insert unsafe type casting at critical points in the 
code, with either "Obj.magic" in Ocaml or "unsafeCoerce" in Haskell.


2) Differences with previous extraction (V6.3 and before)

2.a) The pros 

The ability to extract every Coq term, as explain in the previous 
paragraph. 

The ability to extract from a file an ML module (cf Extraction Library in the 
documentation)

You can have a taste of extraction directly at the toplevel by 
using the "Extraction <ident>" or the "Recursive Extraction <ident>".
This toplevel extraction was already there in V6.3, but was printing 
Fw terms. It now prints in the language of your choice: 
Ocaml, Haskell or Scheme. 

The optimization done on extracted code has been ported between 
V6.3 and V7 and enhanced, and in particular the mechanism of automatic
expansion.

2.b) The cons 

The presence of some parasite "__" as dummy arguments
in functions. This denotes the rests of a proof part. The previous 
extraction was able to remove them totally. The current implementation
removes a good deal of them, but not all. 

This problem is due to extraction upon Type.
For example, let's take this pathological term: 
	(if b then Set else Prop)  : Type
The only way to know if this is an Set (to keep) or a Prop (to remove) 
is to compute the boolean b, and we do not want to do that during 
extraction. 

There is no more "ML import" feature. You can compensate by using 
Axioms, and then "Extract Constant ..." 





[1]:
Exécution de termes de preuves: une nouvelle méthode d'extraction 
pour le Calcul des Constructions Inductives, Pierre Letouzey,
DEA thesis, 2000, 
http://www.pps.jussieu.fr/~letouzey/download/rapport_dea.ps.gz

[2]: 
A New Extraction for Coq, Pierre Letouzey, 
Types 2002 Post-Workshop Proceedings. 
http://www.pps.jussieu.fr/~letouzey/download/extraction2002.ps.gz

[3]: 
Programmation fonctionnelle certifiée: l'extraction de programmes
dans l'assistant Coq. Pierre Letouzey, PhD thesis, 2004.
http://www.pps.jussieu.fr/~letouzey/download/these_letouzey.ps.gz
http://www.pps.jussieu.fr/~letouzey/download/these_letouzey_English.ps.gz

[4]:
Coq Extraction, An overview. Pierre Letouzey. CiE2008.
http://www.pps.jussieu.fr/~letouzey/download/letouzey_extr_cie08.pdf