author | wenzelm |
Sat, 01 Dec 2001 18:52:32 +0100 | |
changeset 12338 | de0f4a63baa5 |
parent 11099 | b301d1f72552 |
child 16417 | 9bc16273c2d4 |
permissions | -rw-r--r-- |
8890 | 1 |
|
9146 | 2 |
header {* Syntactic classes *} |
8890 | 3 |
|
9146 | 4 |
theory Product = Main: |
8890 | 5 |
|
8903 | 6 |
text {* |
12338
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
7 |
\medskip\noindent There is still a feature of Isabelle's type system |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
8 |
left that we have not yet discussed. When declaring polymorphic |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
9 |
constants @{text "c \<Colon> \<sigma>"}, the type variables occurring in @{text \<sigma>} |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
10 |
may be constrained by type classes (or even general sorts) in an |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
11 |
arbitrary way. Note that by default, in Isabelle/HOL the |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
12 |
declaration @{text "\<odot> \<Colon> 'a \<Rightarrow> 'a \<Rightarrow> 'a"} is actually an abbreviation |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
13 |
for @{text "\<odot> \<Colon> 'a\<Colon>type \<Rightarrow> 'a \<Rightarrow> 'a"} Since class @{text type} is the |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
14 |
universal class of HOL, this is not really a constraint at all. |
8903 | 15 |
|
10140 | 16 |
The @{text product} class below provides a less degenerate example of |
8903 | 17 |
syntactic type classes. |
9146 | 18 |
*} |
8903 | 19 |
|
8890 | 20 |
axclass |
12338
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
21 |
product \<subseteq> type |
8890 | 22 |
consts |
10140 | 23 |
product :: "'a\<Colon>product \<Rightarrow> 'a \<Rightarrow> 'a" (infixl "\<odot>" 70) |
8890 | 24 |
|
8903 | 25 |
text {* |
12338
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
26 |
Here class @{text product} is defined as subclass of @{text type} |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
27 |
without any additional axioms. This effects in logical equivalence |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
28 |
of @{text product} and @{text type}, as is reflected by the trivial |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
29 |
introduction rule generated for this definition. |
8903 | 30 |
|
12338
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
31 |
\medskip So what is the difference of declaring @{text "\<odot> \<Colon> |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
32 |
'a\<Colon>product \<Rightarrow> 'a \<Rightarrow> 'a"} vs.\ declaring @{text "\<odot> \<Colon> 'a\<Colon>type \<Rightarrow> 'a \<Rightarrow> |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
33 |
'a"} anyway? In this particular case where @{text "product \<equiv> |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
34 |
type"}, it should be obvious that both declarations are the same |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
35 |
from the logic's point of view. It even makes the most sense to |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
36 |
remove sort constraints from constant declarations, as far as the |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
37 |
purely logical meaning is concerned \cite{Wenzel:1997:TPHOL}. |
8903 | 38 |
|
12338
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
39 |
On the other hand there are syntactic differences, of course. |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
40 |
Constants @{text \<odot>} on some type @{text \<tau>} are rejected by the |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
41 |
type-checker, unless the arity @{text "\<tau> \<Colon> product"} is part of the |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
42 |
type signature. In our example, this arity may be always added when |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
43 |
required by means of an $\INSTANCE$ with the default proof $\DDOT$. |
8903 | 44 |
|
12338
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
45 |
\medskip Thus, we may observe the following discipline of using |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
46 |
syntactic classes. Overloaded polymorphic constants have their type |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
47 |
arguments restricted to an associated (logically trivial) class |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
48 |
@{text c}. Only immediately before \emph{specifying} these |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
49 |
constants on a certain type @{text \<tau>} do we instantiate @{text "\<tau> \<Colon> |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
50 |
c"}. |
8903 | 51 |
|
12338
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
52 |
This is done for class @{text product} and type @{typ bool} as |
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
53 |
follows. |
9146 | 54 |
*} |
8903 | 55 |
|
10309 | 56 |
instance bool :: product .. |
9306 | 57 |
defs (overloaded) |
10140 | 58 |
product_bool_def: "x \<odot> y \<equiv> x \<and> y" |
8890 | 59 |
|
8903 | 60 |
text {* |
10140 | 61 |
The definition @{text prod_bool_def} becomes syntactically |
62 |
well-formed only after the arity @{text "bool \<Colon> product"} is made |
|
63 |
known to the type checker. |
|
8903 | 64 |
|
65 |
\medskip It is very important to see that above $\DEFS$ are not |
|
10223 | 66 |
directly connected with $\INSTANCE$ at all! We were just following |
67 |
our convention to specify @{text \<odot>} on @{typ bool} after having |
|
68 |
instantiated @{text "bool \<Colon> product"}. Isabelle does not require |
|
69 |
these definitions, which is in contrast to programming languages like |
|
70 |
Haskell \cite{haskell-report}. |
|
8903 | 71 |
|
72 |
\medskip While Isabelle type classes and those of Haskell are almost |
|
73 |
the same as far as type-checking and type inference are concerned, |
|
8907 | 74 |
there are important semantic differences. Haskell classes require |
75 |
their instances to \emph{provide operations} of certain \emph{names}. |
|
8903 | 76 |
Therefore, its \texttt{instance} has a \texttt{where} part that tells |
77 |
the system what these ``member functions'' should be. |
|
78 |
||
10140 | 79 |
This style of \texttt{instance} would not make much sense in |
80 |
Isabelle's meta-logic, because there is no internal notion of |
|
81 |
``providing operations'' or even ``names of functions''. |
|
9146 | 82 |
*} |
8903 | 83 |
|
12338
de0f4a63baa5
renamed class "term" to "type" (actually "HOL.type");
wenzelm
parents:
11099
diff
changeset
|
84 |
end |