wenzelm@7046
|
1 |
|
wenzelm@13048
|
2 |
\chapter{Syntax primitives}
|
wenzelm@7046
|
3 |
|
wenzelm@12618
|
4 |
The rather generic framework of Isabelle/Isar syntax emerges from three main
|
wenzelm@12618
|
5 |
syntactic categories: \emph{commands} of the top-level Isar engine (covering
|
wenzelm@12618
|
6 |
theory and proof elements), \emph{methods} for general goal refinements
|
wenzelm@12618
|
7 |
(analogous to traditional ``tactics''), and \emph{attributes} for operations
|
wenzelm@12618
|
8 |
on facts (within a certain context). Here we give a reference of basic
|
wenzelm@12618
|
9 |
syntactic entities underlying Isabelle/Isar syntax in a bottom-up manner.
|
wenzelm@12618
|
10 |
Concrete theory and proof language elements will be introduced later on.
|
wenzelm@7134
|
11 |
|
wenzelm@7315
|
12 |
\medskip
|
wenzelm@7046
|
13 |
|
wenzelm@7315
|
14 |
In order to get started with writing well-formed Isabelle/Isar documents, the
|
wenzelm@7315
|
15 |
most important aspect to be noted is the difference of \emph{inner} versus
|
wenzelm@7315
|
16 |
\emph{outer} syntax. Inner syntax is that of Isabelle types and terms of the
|
wenzelm@12618
|
17 |
logic, while outer syntax is that of Isabelle/Isar theory sources (including
|
wenzelm@7895
|
18 |
proofs). As a general rule, inner syntax entities may occur only as
|
wenzelm@7895
|
19 |
\emph{atomic entities} within outer syntax. For example, the string
|
wenzelm@7895
|
20 |
\texttt{"x + y"} and identifier \texttt{z} are legal term specifications
|
wenzelm@7895
|
21 |
within a theory, while \texttt{x + y} is not.
|
wenzelm@7315
|
22 |
|
wenzelm@7315
|
23 |
\begin{warn}
|
wenzelm@12618
|
24 |
Old-style Isabelle theories used to fake parts of the inner syntax of types,
|
wenzelm@12618
|
25 |
with rather complicated rules when quotes may be omitted. Despite the minor
|
wenzelm@12618
|
26 |
drawback of requiring quotes more often, the syntax of Isabelle/Isar is
|
wenzelm@12618
|
27 |
somewhat simpler and more robust in that respect.
|
wenzelm@7315
|
28 |
\end{warn}
|
wenzelm@7315
|
29 |
|
wenzelm@12618
|
30 |
Printed theory documents usually omit quotes to gain readability (this is a
|
wenzelm@12618
|
31 |
matter of {\LaTeX} macro setup, say via \verb,\isabellestyle,, see also
|
wenzelm@12618
|
32 |
\cite{isabelle-sys}). Experienced users of Isabelle/Isar may easily
|
wenzelm@12618
|
33 |
reconstruct the lost technical information, while mere readers need not care
|
wenzelm@12618
|
34 |
about quotes at all.
|
wenzelm@12618
|
35 |
|
wenzelm@7466
|
36 |
\medskip
|
wenzelm@7466
|
37 |
|
wenzelm@9601
|
38 |
Isabelle/Isar input may contain any number of input termination characters
|
wenzelm@12618
|
39 |
``\texttt{;}'' (semicolon) to separate commands explicitly. This is
|
wenzelm@9601
|
40 |
particularly useful in interactive shell sessions to make clear where the
|
wenzelm@12618
|
41 |
current command is intended to end. Otherwise, the interpreter loop will
|
wenzelm@12618
|
42 |
continue to issue a secondary prompt ``\verb,#,'' until an end-of-command is
|
wenzelm@13039
|
43 |
clearly recognized from the input syntax, e.g.\ encounter of the next command
|
wenzelm@12618
|
44 |
keyword.
|
wenzelm@9601
|
45 |
|
wenzelm@9601
|
46 |
Advanced interfaces such as Proof~General \cite{proofgeneral} do not require
|
wenzelm@9601
|
47 |
explicit semicolons, the amount of input text is determined automatically by
|
wenzelm@12618
|
48 |
inspecting the present content of the Emacs text buffer. In the printed
|
wenzelm@12618
|
49 |
presentation of Isabelle/Isar documents semicolons are omitted altogether for
|
wenzelm@7981
|
50 |
readability.
|
wenzelm@7466
|
51 |
|
wenzelm@12618
|
52 |
\begin{warn}
|
wenzelm@12618
|
53 |
Proof~General requires certain syntax classification tables in order to
|
wenzelm@12618
|
54 |
achieve properly synchronized interaction with the Isabelle/Isar process.
|
wenzelm@12618
|
55 |
These tables need to be consistent with the Isabelle version and particular
|
wenzelm@12618
|
56 |
logic image to be used in a running session (common object-logics may well
|
wenzelm@12618
|
57 |
change the outer syntax). The standard setup should work correctly with any
|
wenzelm@12618
|
58 |
of the ``official'' logic images derived from Isabelle/HOL (including HOLCF
|
wenzelm@12618
|
59 |
etc.). Users of alternative logics may need to tell Proof~General
|
wenzelm@12618
|
60 |
explicitly, e.g.\ by giving an option \verb,-k ZF, (in conjunction with
|
wenzelm@12618
|
61 |
\verb,-l ZF, to specify the default logic image).
|
wenzelm@12618
|
62 |
\end{warn}
|
wenzelm@7315
|
63 |
|
wenzelm@7315
|
64 |
\section{Lexical matters}\label{sec:lex-syntax}
|
wenzelm@7315
|
65 |
|
wenzelm@14955
|
66 |
The Isabelle/Isar outer syntax provides token classes as presented below; most
|
wenzelm@14955
|
67 |
of these coincide with the inner lexical syntax as presented in
|
wenzelm@14955
|
68 |
\cite{isabelle-ref}.
|
wenzelm@7315
|
69 |
|
wenzelm@9617
|
70 |
\indexoutertoken{ident}\indexoutertoken{longident}\indexoutertoken{symident}
|
wenzelm@9617
|
71 |
\indexoutertoken{nat}\indexoutertoken{var}\indexoutertoken{typefree}
|
wenzelm@9617
|
72 |
\indexoutertoken{typevar}\indexoutertoken{string}\indexoutertoken{verbatim}
|
wenzelm@7315
|
73 |
\begin{matharray}{rcl}
|
wenzelm@14955
|
74 |
ident & = & letter\,quasiletter^* \\
|
wenzelm@14955
|
75 |
longident & = & ident (\verb,.,ident)^+ \\
|
wenzelm@14955
|
76 |
symident & = & sym^+ ~|~ \verb,\<,ident\verb,>, \\
|
wenzelm@7315
|
77 |
nat & = & digit^+ \\
|
ballarin@14212
|
78 |
var & = & ident ~|~ \verb,?,ident ~|~ \verb,?,ident\verb,.,nat \\
|
wenzelm@7315
|
79 |
typefree & = & \verb,',ident \\
|
ballarin@14212
|
80 |
typevar & = & typefree ~|~ \verb,?,typefree ~|~ \verb,?,typefree\verb,.,nat \\
|
wenzelm@7315
|
81 |
string & = & \verb,", ~\dots~ \verb,", \\
|
kleing@14483
|
82 |
verbatim & = & \verb,{*, ~\dots~ \verb,*}, \\[1ex]
|
kleing@14483
|
83 |
|
wenzelm@14960
|
84 |
letter & = & latin ~|~ \verb,\<,latin\verb,>, ~|~ \verb,\<,latin\,latin\verb,>, ~|~ greek ~|~ \\
|
wenzelm@14960
|
85 |
& & \verb,\<^isub>, ~|~ \verb,\<^isup>, \\
|
wenzelm@14960
|
86 |
quasiletter & = & letter ~|~ digit ~|~ \verb,_, ~|~ \verb,', \\
|
wenzelm@14895
|
87 |
latin & = & \verb,a, ~|~ \dots ~|~ \verb,z, ~|~ \verb,A, ~|~ \dots ~|~ \verb,Z, \\
|
wenzelm@7315
|
88 |
digit & = & \verb,0, ~|~ \dots ~|~ \verb,9, \\
|
wenzelm@7315
|
89 |
sym & = & \verb,!, ~|~ \verb,#, ~|~ \verb,$, ~|~ \verb,%, ~|~ \verb,&, ~|~ %$
|
wenzelm@10160
|
90 |
\verb,*, ~|~ \verb,+, ~|~ \verb,-, ~|~ \verb,/, ~|~ \verb,:, ~|~ \\
|
wenzelm@10160
|
91 |
& & \verb,<, ~|~ \verb,=, ~|~ \verb,>, ~|~ \verb,?, ~|~ \texttt{\at} ~|~
|
wenzelm@10160
|
92 |
\verb,^, ~|~ \verb,_, ~|~ \verb,`, ~|~ \verb,|, ~|~ \verb,~, \\
|
wenzelm@14955
|
93 |
greek & = & \verb,\<alpha>, ~|~ \verb,\<beta>, ~|~ \verb,\<gamma>, ~|~ \verb,\<delta>, ~| \\
|
wenzelm@14955
|
94 |
& & \verb,\<epsilon>, ~|~ \verb,\<zeta>, ~|~ \verb,\<eta>, ~|~ \verb,\<theta>, ~| \\
|
wenzelm@14955
|
95 |
& & \verb,\<iota>, ~|~ \verb,\<kappa>, ~|~ \verb,\<mu>, ~|~ \verb,\<nu>, ~| \\
|
wenzelm@14955
|
96 |
& & \verb,\<xi>, ~|~ \verb,\<pi>, ~|~ \verb,\<rho>, ~|~ \verb,\<sigma>, ~| \\
|
wenzelm@14955
|
97 |
& & \verb,\<tau>, ~|~ \verb,\<upsilon>, ~|~ \verb,\<phi>, ~|~ \verb,\<psi>, ~| \\
|
wenzelm@14955
|
98 |
& & \verb,\<omega>, ~|~ \verb,\<Gamma>, ~|~ \verb,\<Delta>, ~|~ \verb,\<Theta>, ~| \\
|
wenzelm@14955
|
99 |
& & \verb,\<Lambda>, ~|~ \verb,\<Xi>, ~|~ \verb,\<Pi>, ~|~ \verb,\<Sigma>, ~| \\
|
wenzelm@14955
|
100 |
& & \verb,\<Upsilon>, ~|~ \verb,\<Phi>, ~|~ \verb,\<Psi>, ~|~ \verb,\<Omega>, \\
|
wenzelm@7315
|
101 |
\end{matharray}
|
wenzelm@7315
|
102 |
|
wenzelm@13039
|
103 |
The syntax of $string$ admits any characters, including newlines; ``\verb|"|''
|
wenzelm@13039
|
104 |
(double-quote) and ``\verb|\|'' (backslash) need to be escaped by a backslash.
|
wenzelm@14955
|
105 |
The body of $verbatim$ may consist of any text not containing ``\verb|*}|'';
|
wenzelm@14955
|
106 |
this allows convenient inclusion of quotes without further escapes. The greek
|
wenzelm@14955
|
107 |
letters do \emph{not} include \verb,\<lambda>,, which is already used
|
wenzelm@14955
|
108 |
differently in the meta-logic.
|
wenzelm@7315
|
109 |
|
wenzelm@14955
|
110 |
Common mathematical symbols such as $\forall$ are represented in Isabelle as
|
wenzelm@14955
|
111 |
\verb,\<forall>,. There are infinitely many legal symbols like this, although
|
wenzelm@14955
|
112 |
proper presentation is left to front-end tools such as {\LaTeX} or
|
wenzelm@14955
|
113 |
Proof~General with the X-Symbol package. A list of standard Isabelle symbols
|
wenzelm@14955
|
114 |
that work well with these tools is given in \cite[appendix~A]{isabelle-sys}.
|
wenzelm@14955
|
115 |
|
wenzelm@14955
|
116 |
Comments take the form \texttt{(*~\dots~*)} and may be nested, although
|
wenzelm@14955
|
117 |
user-interface tools may prevent this. Note that \texttt{(*~\dots~*)}
|
wenzelm@14955
|
118 |
indicate source comments only, which are stripped after lexical analysis of
|
wenzelm@14955
|
119 |
the input. The Isar document syntax also provides formal comments that are
|
wenzelm@14955
|
120 |
considered as part of the text (see \S\ref{sec:comments}).
|
wenzelm@12618
|
121 |
|
wenzelm@12618
|
122 |
\begin{warn}
|
wenzelm@12618
|
123 |
Proof~General does not handle nested comments properly; it is also unable to
|
wenzelm@12618
|
124 |
keep \verb,(*,\,/\,\verb,{*, and \verb,*),\,/\,\verb,*}, apart, despite
|
wenzelm@12618
|
125 |
their rather different meaning. These are inherent problems of Emacs
|
wenzelm@13039
|
126 |
legacy. Users should not be overly aggressive about nesting or alternating
|
wenzelm@13039
|
127 |
these delimiters.
|
wenzelm@12618
|
128 |
\end{warn}
|
wenzelm@12618
|
129 |
|
wenzelm@7046
|
130 |
|
wenzelm@7046
|
131 |
\section{Common syntax entities}
|
wenzelm@7046
|
132 |
|
wenzelm@7335
|
133 |
Subsequently, we introduce several basic syntactic entities, such as names,
|
wenzelm@7895
|
134 |
terms, and theorem specifications, which have been factored out of the actual
|
wenzelm@7895
|
135 |
Isar language elements to be described later.
|
wenzelm@7134
|
136 |
|
wenzelm@14895
|
137 |
Note that some of the basic syntactic entities introduced below (e.g.\
|
wenzelm@14895
|
138 |
\railqtok{name}) act much like tokens rather than plain nonterminals (e.g.\
|
wenzelm@7895
|
139 |
\railnonterm{sort}), especially for the sake of error messages. E.g.\ syntax
|
wenzelm@13048
|
140 |
elements like $\CONSTS$ referring to \railqtok{name} or \railqtok{type} would
|
wenzelm@13048
|
141 |
really report a missing name or type rather than any of the constituent
|
wenzelm@13048
|
142 |
primitive tokens such as \railtok{ident} or \railtok{string}.
|
wenzelm@7046
|
143 |
|
wenzelm@7050
|
144 |
|
wenzelm@7050
|
145 |
\subsection{Names}
|
wenzelm@7050
|
146 |
|
wenzelm@13048
|
147 |
Entity \railqtok{name} usually refers to any name of types, constants,
|
wenzelm@7167
|
148 |
theorems etc.\ that are to be \emph{declared} or \emph{defined} (so qualified
|
wenzelm@8548
|
149 |
identifiers are excluded here). Quoted strings provide an escape for
|
wenzelm@14895
|
150 |
non-identifier names or those ruled out by outer syntax keywords (e.g.\
|
wenzelm@7134
|
151 |
\verb|"let"|). Already existing objects are usually referenced by
|
wenzelm@13048
|
152 |
\railqtok{nameref}.
|
wenzelm@7050
|
153 |
|
wenzelm@7141
|
154 |
\indexoutertoken{name}\indexoutertoken{parname}\indexoutertoken{nameref}
|
wenzelm@9617
|
155 |
\indexoutertoken{int}
|
wenzelm@7046
|
156 |
\begin{rail}
|
wenzelm@8145
|
157 |
name: ident | symident | string | nat
|
wenzelm@7046
|
158 |
;
|
wenzelm@7167
|
159 |
parname: '(' name ')'
|
wenzelm@7141
|
160 |
;
|
wenzelm@7167
|
161 |
nameref: name | longident
|
wenzelm@7046
|
162 |
;
|
wenzelm@9617
|
163 |
int: nat | '-' nat
|
wenzelm@9617
|
164 |
;
|
wenzelm@7046
|
165 |
\end{rail}
|
wenzelm@7046
|
166 |
|
wenzelm@7050
|
167 |
|
wenzelm@7315
|
168 |
\subsection{Comments}\label{sec:comments}
|
wenzelm@7046
|
169 |
|
wenzelm@13048
|
170 |
Large chunks of plain \railqtok{text} are usually given \railtok{verbatim},
|
wenzelm@13048
|
171 |
i.e.\ enclosed in \verb|{*|~\dots~\verb|*}|. For convenience, any of the
|
wenzelm@13048
|
172 |
smaller text units conforming to \railqtok{nameref} are admitted as well. A
|
wenzelm@13048
|
173 |
marginal \railnonterm{comment} is of the form \texttt{--} \railqtok{text}.
|
wenzelm@13048
|
174 |
Any number of these may occur within Isabelle/Isar commands.
|
wenzelm@12618
|
175 |
|
wenzelm@12879
|
176 |
\indexoutertoken{text}\indexouternonterm{comment}
|
wenzelm@7046
|
177 |
\begin{rail}
|
wenzelm@7167
|
178 |
text: verbatim | nameref
|
wenzelm@7050
|
179 |
;
|
wenzelm@12879
|
180 |
comment: '--' text
|
wenzelm@7046
|
181 |
;
|
wenzelm@7046
|
182 |
\end{rail}
|
wenzelm@7046
|
183 |
|
wenzelm@7046
|
184 |
|
wenzelm@7335
|
185 |
\subsection{Type classes, sorts and arities}
|
wenzelm@7046
|
186 |
|
wenzelm@8896
|
187 |
Classes are specified by plain names. Sorts have a very simple inner syntax,
|
wenzelm@8896
|
188 |
which is either a single class name $c$ or a list $\{c@1, \dots, c@n\}$
|
wenzelm@8896
|
189 |
referring to the intersection of these classes. The syntax of type arities is
|
wenzelm@8896
|
190 |
given directly at the outer level.
|
wenzelm@7050
|
191 |
|
wenzelm@11100
|
192 |
\railalias{subseteq}{\isasymsubseteq}
|
wenzelm@11100
|
193 |
\railterm{subseteq}
|
wenzelm@11100
|
194 |
|
wenzelm@14605
|
195 |
\indexouternonterm{sort}\indexouternonterm{arity}
|
wenzelm@7135
|
196 |
\indexouternonterm{classdecl}
|
wenzelm@7046
|
197 |
\begin{rail}
|
wenzelm@11100
|
198 |
classdecl: name (('<' | subseteq) (nameref + ','))?
|
wenzelm@7046
|
199 |
;
|
wenzelm@8896
|
200 |
sort: nameref
|
wenzelm@7046
|
201 |
;
|
wenzelm@7167
|
202 |
arity: ('(' (sort + ',') ')')? sort
|
wenzelm@7046
|
203 |
;
|
wenzelm@7046
|
204 |
\end{rail}
|
wenzelm@7046
|
205 |
|
wenzelm@7046
|
206 |
|
wenzelm@7167
|
207 |
\subsection{Types and terms}\label{sec:types-terms}
|
wenzelm@7046
|
208 |
|
wenzelm@7167
|
209 |
The actual inner Isabelle syntax, that of types and terms of the logic, is far
|
wenzelm@7895
|
210 |
too sophisticated in order to be modelled explicitly at the outer theory
|
wenzelm@8548
|
211 |
level. Basically, any such entity has to be quoted to turn it into a single
|
wenzelm@8548
|
212 |
token (the parsing and type-checking is performed internally later). For
|
wenzelm@8548
|
213 |
convenience, a slightly more liberal convention is adopted: quotes may be
|
wenzelm@13039
|
214 |
omitted for any type or term that is already atomic at the outer level. For
|
wenzelm@13039
|
215 |
example, one may just write \texttt{x} instead of \texttt{"x"}. Note that
|
wenzelm@8548
|
216 |
symbolic identifiers (e.g.\ \texttt{++} or $\forall$) are available as well,
|
wenzelm@12618
|
217 |
provided these have not been superseded by commands or other keywords already
|
wenzelm@12618
|
218 |
(e.g.\ \texttt{=} or \texttt{+}).
|
wenzelm@7050
|
219 |
|
wenzelm@7050
|
220 |
\indexoutertoken{type}\indexoutertoken{term}\indexoutertoken{prop}
|
wenzelm@7046
|
221 |
\begin{rail}
|
wenzelm@7167
|
222 |
type: nameref | typefree | typevar
|
wenzelm@7050
|
223 |
;
|
wenzelm@8593
|
224 |
term: nameref | var
|
wenzelm@7050
|
225 |
;
|
wenzelm@7167
|
226 |
prop: term
|
wenzelm@7050
|
227 |
;
|
wenzelm@7046
|
228 |
\end{rail}
|
wenzelm@7046
|
229 |
|
wenzelm@8690
|
230 |
Positional instantiations are indicated by giving a sequence of terms, or the
|
wenzelm@8690
|
231 |
placeholder ``$\_$'' (underscore), which means to skip a position.
|
wenzelm@8690
|
232 |
|
wenzelm@8690
|
233 |
\indexoutertoken{inst}\indexoutertoken{insts}
|
wenzelm@8690
|
234 |
\begin{rail}
|
wenzelm@8690
|
235 |
inst: underscore | term
|
wenzelm@8690
|
236 |
;
|
wenzelm@8690
|
237 |
insts: (inst *)
|
wenzelm@8690
|
238 |
;
|
wenzelm@8690
|
239 |
\end{rail}
|
wenzelm@8690
|
240 |
|
wenzelm@7167
|
241 |
Type declarations and definitions usually refer to \railnonterm{typespec} on
|
wenzelm@7167
|
242 |
the left-hand side. This models basic type constructor application at the
|
wenzelm@7167
|
243 |
outer syntax level. Note that only plain postfix notation is available here,
|
wenzelm@7167
|
244 |
but no infixes.
|
wenzelm@7050
|
245 |
|
wenzelm@7050
|
246 |
\indexouternonterm{typespec}
|
wenzelm@7050
|
247 |
\begin{rail}
|
wenzelm@7167
|
248 |
typespec: (() | typefree | '(' ( typefree + ',' ) ')') name
|
wenzelm@7050
|
249 |
;
|
wenzelm@7050
|
250 |
\end{rail}
|
wenzelm@7050
|
251 |
|
wenzelm@7050
|
252 |
|
wenzelm@7046
|
253 |
\subsection{Mixfix annotations}
|
wenzelm@7046
|
254 |
|
wenzelm@7134
|
255 |
Mixfix annotations specify concrete \emph{inner} syntax of Isabelle types and
|
wenzelm@12618
|
256 |
terms. Some commands such as $\TYPES$ (see \S\ref{sec:types-pure}) admit
|
wenzelm@12618
|
257 |
infixes only, while $\CONSTS$ (see \S\ref{sec:consts}) and
|
wenzelm@12618
|
258 |
$\isarkeyword{syntax}$ (see \S\ref{sec:syn-trans}) support the full range of
|
wenzelm@12618
|
259 |
general mixfixes and binders.
|
wenzelm@7046
|
260 |
|
wenzelm@12976
|
261 |
\indexouternonterm{infix}\indexouternonterm{mixfix}\indexouternonterm{structmixfix}
|
wenzelm@7050
|
262 |
\begin{rail}
|
wenzelm@11651
|
263 |
infix: '(' ('infix' | 'infixl' | 'infixr') string? nat ')'
|
wenzelm@7167
|
264 |
;
|
wenzelm@7175
|
265 |
mixfix: infix | '(' string prios? nat? ')' | '(' 'binder' string prios? nat ')'
|
wenzelm@7050
|
266 |
;
|
wenzelm@12976
|
267 |
structmixfix: mixfix | '(' 'structure' ')'
|
wenzelm@12976
|
268 |
;
|
wenzelm@7046
|
269 |
|
wenzelm@7175
|
270 |
prios: '[' (nat + ',') ']'
|
wenzelm@7050
|
271 |
;
|
wenzelm@7050
|
272 |
\end{rail}
|
wenzelm@7046
|
273 |
|
wenzelm@13048
|
274 |
Here the \railtok{string} specifications refer to the actual mixfix template
|
wenzelm@12618
|
275 |
(see also \cite{isabelle-ref}), which may include literal text, spacing,
|
wenzelm@12618
|
276 |
blocks, and arguments (denoted by ``$_$''); the special symbol \verb,\<index>,
|
wenzelm@12618
|
277 |
(printed as ``\i'') represents an index argument that specifies an implicit
|
wenzelm@12618
|
278 |
structure reference (see also \S\ref{sec:locale}). Infix and binder
|
wenzelm@12618
|
279 |
declarations provide common abbreviations for particular mixfix declarations.
|
wenzelm@12618
|
280 |
So in practice, mixfix templates mostly degenerate to literal text for
|
wenzelm@12618
|
281 |
concrete syntax, such as ``\verb,++,'' for an infix symbol, or ``\verb,++,\i''
|
wenzelm@12618
|
282 |
for an infix of an implicit structure.
|
wenzelm@12618
|
283 |
|
wenzelm@12618
|
284 |
|
wenzelm@12618
|
285 |
|
wenzelm@12618
|
286 |
\subsection{Proof methods}\label{sec:syn-meth}
|
wenzelm@12618
|
287 |
|
wenzelm@12618
|
288 |
Proof methods are either basic ones, or expressions composed of methods via
|
wenzelm@12618
|
289 |
``\texttt{,}'' (sequential composition), ``\texttt{|}'' (alternative choices),
|
wenzelm@12618
|
290 |
``\texttt{?}'' (try), ``\texttt{+}'' (repeat at least once). In practice,
|
wenzelm@12618
|
291 |
proof methods are usually just a comma separated list of
|
wenzelm@13048
|
292 |
\railqtok{nameref}~\railnonterm{args} specifications. Note that parentheses
|
wenzelm@12618
|
293 |
may be dropped for single method specifications (with no arguments).
|
wenzelm@12618
|
294 |
|
wenzelm@12618
|
295 |
\indexouternonterm{method}
|
wenzelm@12618
|
296 |
\begin{rail}
|
wenzelm@12618
|
297 |
method: (nameref | '(' methods ')') (() | '?' | '+')
|
wenzelm@12618
|
298 |
;
|
wenzelm@12618
|
299 |
methods: (nameref args | method) + (',' | '|')
|
wenzelm@12618
|
300 |
;
|
wenzelm@12618
|
301 |
\end{rail}
|
wenzelm@12618
|
302 |
|
wenzelm@12618
|
303 |
Proper use of Isar proof methods does \emph{not} involve goal addressing.
|
wenzelm@12618
|
304 |
Nevertheless, specifying goal ranges may occasionally come in handy in
|
wenzelm@12618
|
305 |
emulating tactic scripts. Note that $[n-]$ refers to all goals, starting from
|
wenzelm@12618
|
306 |
$n$. All goals may be specified by $[!]$, which is the same as $[1-]$.
|
wenzelm@12618
|
307 |
|
wenzelm@12618
|
308 |
\indexouternonterm{goalspec}
|
wenzelm@12618
|
309 |
\begin{rail}
|
wenzelm@12618
|
310 |
goalspec: '[' (nat '-' nat | nat '-' | nat | '!' ) ']'
|
wenzelm@12618
|
311 |
;
|
wenzelm@12618
|
312 |
\end{rail}
|
wenzelm@12618
|
313 |
|
wenzelm@7050
|
314 |
|
wenzelm@7134
|
315 |
\subsection{Attributes and theorems}\label{sec:syn-att}
|
wenzelm@7050
|
316 |
|
wenzelm@7050
|
317 |
Attributes (and proof methods, see \S\ref{sec:syn-meth}) have their own
|
wenzelm@7335
|
318 |
``semi-inner'' syntax, in the sense that input conforming to
|
wenzelm@7335
|
319 |
\railnonterm{args} below is parsed by the attribute a second time. The
|
wenzelm@7335
|
320 |
attribute argument specifications may be any sequence of atomic entities
|
wenzelm@7335
|
321 |
(identifiers, strings etc.), or properly bracketed argument lists. Below
|
wenzelm@13048
|
322 |
\railqtok{atom} refers to any atomic entity, including any \railtok{keyword}
|
wenzelm@13048
|
323 |
conforming to \railtok{symident}.
|
wenzelm@7050
|
324 |
|
wenzelm@7050
|
325 |
\indexoutertoken{atom}\indexouternonterm{args}\indexouternonterm{attributes}
|
wenzelm@7050
|
326 |
\begin{rail}
|
wenzelm@7466
|
327 |
atom: nameref | typefree | typevar | var | nat | keyword
|
wenzelm@7050
|
328 |
;
|
wenzelm@8896
|
329 |
arg: atom | '(' args ')' | '[' args ']'
|
wenzelm@7134
|
330 |
;
|
wenzelm@7167
|
331 |
args: arg *
|
wenzelm@7134
|
332 |
;
|
wenzelm@7167
|
333 |
attributes: '[' (nameref args * ',') ']'
|
wenzelm@7050
|
334 |
;
|
wenzelm@7050
|
335 |
\end{rail}
|
wenzelm@7050
|
336 |
|
wenzelm@7895
|
337 |
Theorem specifications come in several flavors: \railnonterm{axmdecl} and
|
wenzelm@7175
|
338 |
\railnonterm{thmdecl} usually refer to axioms, assumptions or results of goal
|
wenzelm@7981
|
339 |
statements, while \railnonterm{thmdef} collects lists of existing theorems.
|
wenzelm@7981
|
340 |
Existing theorems are given by \railnonterm{thmref} and \railnonterm{thmrefs},
|
wenzelm@7981
|
341 |
the former requires an actual singleton result. Any of these theorem
|
wenzelm@7175
|
342 |
specifications may include lists of attributes both on the left and right hand
|
wenzelm@13039
|
343 |
sides; attributes are applied to any immediately preceding fact. If names are
|
wenzelm@13039
|
344 |
omitted, the theorems are not stored within the theorem database of the theory
|
wenzelm@13039
|
345 |
or proof context; any given attributes are still applied, though.
|
wenzelm@7050
|
346 |
|
wenzelm@7135
|
347 |
\indexouternonterm{thmdecl}\indexouternonterm{axmdecl}
|
wenzelm@7135
|
348 |
\indexouternonterm{thmdef}\indexouternonterm{thmrefs}
|
wenzelm@7050
|
349 |
\begin{rail}
|
wenzelm@7167
|
350 |
axmdecl: name attributes? ':'
|
wenzelm@7050
|
351 |
;
|
wenzelm@9200
|
352 |
thmdecl: thmbind ':'
|
wenzelm@7135
|
353 |
;
|
wenzelm@9200
|
354 |
thmdef: thmbind '='
|
wenzelm@7050
|
355 |
;
|
wenzelm@7175
|
356 |
thmref: nameref attributes?
|
wenzelm@7175
|
357 |
;
|
wenzelm@7175
|
358 |
thmrefs: thmref +
|
wenzelm@7134
|
359 |
;
|
wenzelm@7167
|
360 |
|
wenzelm@9200
|
361 |
thmbind: name attributes | name | attributes
|
wenzelm@7050
|
362 |
;
|
wenzelm@7050
|
363 |
\end{rail}
|
wenzelm@7050
|
364 |
|
wenzelm@7050
|
365 |
|
wenzelm@12618
|
366 |
\subsection{Term patterns and declarations}\label{sec:term-decls}
|
wenzelm@7050
|
367 |
|
wenzelm@12618
|
368 |
Wherever explicit propositions (or term fragments) occur in a proof text,
|
wenzelm@12618
|
369 |
casual binding of schematic term variables may be given specified via patterns
|
wenzelm@13039
|
370 |
of the form ``$\ISS{p@1\;\dots}{p@n}$''. There are separate versions
|
wenzelm@13048
|
371 |
available for \railqtok{term}s and \railqtok{prop}s. The latter provides a
|
wenzelm@13048
|
372 |
$\CONCLNAME$ part with patterns referring the (atomic) conclusion of a rule.
|
wenzelm@7050
|
373 |
|
wenzelm@12618
|
374 |
\indexouternonterm{termpat}\indexouternonterm{proppat}
|
wenzelm@7050
|
375 |
\begin{rail}
|
wenzelm@12618
|
376 |
termpat: '(' ('is' term +) ')'
|
wenzelm@7134
|
377 |
;
|
wenzelm@12618
|
378 |
proppat: '(' (('is' prop +) | 'concl' ('is' prop +) | ('is' prop +) 'concl' ('is' prop +)) ')'
|
wenzelm@7050
|
379 |
;
|
wenzelm@7050
|
380 |
\end{rail}
|
wenzelm@7046
|
381 |
|
wenzelm@12618
|
382 |
Declarations of local variables $x :: \tau$ and logical propositions $a :
|
wenzelm@12618
|
383 |
\phi$ represent different views on the same principle of introducing a local
|
wenzelm@12618
|
384 |
scope. In practice, one may usually omit the typing of $vars$ (due to
|
wenzelm@13039
|
385 |
type-inference), and the naming of propositions (due to implicit references of
|
wenzelm@13039
|
386 |
current facts). In any case, Isar proof elements usually admit to introduce
|
wenzelm@12618
|
387 |
multiple such items simultaneously.
|
wenzelm@8532
|
388 |
|
wenzelm@12618
|
389 |
\indexouternonterm{vars}\indexouternonterm{props}
|
wenzelm@8532
|
390 |
\begin{rail}
|
wenzelm@12618
|
391 |
vars: (name+) ('::' type)?
|
wenzelm@12618
|
392 |
;
|
wenzelm@12618
|
393 |
props: thmdecl? (prop proppat? +)
|
wenzelm@8532
|
394 |
;
|
wenzelm@8532
|
395 |
\end{rail}
|
wenzelm@8532
|
396 |
|
wenzelm@12618
|
397 |
The treatment of multiple declarations corresponds to the complementary focus
|
wenzelm@12618
|
398 |
of $vars$ versus $props$: in ``$x@1~\dots~x@n :: \tau$'' the typing refers to
|
wenzelm@12618
|
399 |
all variables, while in $a\colon \phi@1~\dots~\phi@n$ the naming refers to all
|
wenzelm@12618
|
400 |
propositions collectively. Isar language elements that refer to $vars$ or
|
wenzelm@12618
|
401 |
$props$ typically admit separate typings or namings via another level of
|
wenzelm@12618
|
402 |
iteration, with explicit $\AND$ separators; e.g.\ see $\FIXNAME$ and
|
wenzelm@12618
|
403 |
$\ASSUMENAME$ in \S\ref{sec:proof-context}.
|
wenzelm@12618
|
404 |
|
wenzelm@7046
|
405 |
|
wenzelm@9200
|
406 |
\subsection{Antiquotations}\label{sec:antiq}
|
wenzelm@9200
|
407 |
|
wenzelm@10336
|
408 |
\begin{matharray}{rcl}
|
wenzelm@10336
|
409 |
thm & : & \isarantiq \\
|
wenzelm@10336
|
410 |
prop & : & \isarantiq \\
|
wenzelm@10336
|
411 |
term & : & \isarantiq \\
|
wenzelm@10336
|
412 |
typ & : & \isarantiq \\
|
wenzelm@10336
|
413 |
text & : & \isarantiq \\
|
wenzelm@10336
|
414 |
goals & : & \isarantiq \\
|
nipkow@10351
|
415 |
subgoals & : & \isarantiq \\
|
berghofe@13827
|
416 |
prf & : & \isarantiq \\
|
berghofe@13827
|
417 |
full_prf & : & \isarantiq \\
|
wenzelm@10336
|
418 |
\end{matharray}
|
wenzelm@10336
|
419 |
|
wenzelm@9200
|
420 |
The text body of formal comments (see also \S\ref{sec:comments}) may contain
|
wenzelm@9200
|
421 |
antiquotations of logical entities, such as theorems, terms and types, which
|
wenzelm@9200
|
422 |
are to be presented in the final output produced by the Isabelle document
|
wenzelm@9200
|
423 |
preparation system (see also \S\ref{sec:document-prep}).
|
wenzelm@9200
|
424 |
|
wenzelm@9601
|
425 |
Thus embedding of
|
wenzelm@13039
|
426 |
``\texttt{{\at}{\ttlbrace}term~[show_types]~"f(x)~=~a~+~x"{\ttrbrace}}''
|
wenzelm@13039
|
427 |
within a text block would cause
|
wenzelm@9200
|
428 |
\isa{(f{\isasymColon}'a~{\isasymRightarrow}~'a)~(x{\isasymColon}'a)~=~(a{\isasymColon}'a)~+~x}
|
wenzelm@10160
|
429 |
to appear in the final {\LaTeX} document. Also note that theorem
|
wenzelm@10160
|
430 |
antiquotations may involve attributes as well. For example,
|
wenzelm@10160
|
431 |
\texttt{{\at}{\ttlbrace}thm~sym~[no_vars]{\ttrbrace}} would print the
|
wenzelm@10160
|
432 |
statement where all schematic variables have been replaced by fixed ones,
|
wenzelm@12618
|
433 |
which are easier to read.
|
wenzelm@9200
|
434 |
|
wenzelm@9728
|
435 |
\indexisarant{thm}\indexisarant{prop}\indexisarant{term}
|
wenzelm@10355
|
436 |
\indexisarant{typ}\indexisarant{text}\indexisarant{goals}\indexisarant{subgoals}
|
wenzelm@9200
|
437 |
\begin{rail}
|
wenzelm@9200
|
438 |
atsign lbrace antiquotation rbrace
|
wenzelm@9200
|
439 |
;
|
wenzelm@9200
|
440 |
|
wenzelm@9200
|
441 |
antiquotation:
|
wenzelm@9200
|
442 |
'thm' options thmrefs |
|
wenzelm@9200
|
443 |
'prop' options prop |
|
wenzelm@9200
|
444 |
'term' options term |
|
wenzelm@9728
|
445 |
'typ' options type |
|
wenzelm@10319
|
446 |
'text' options name |
|
wenzelm@10355
|
447 |
'goals' options |
|
berghofe@13827
|
448 |
'subgoals' options |
|
berghofe@13827
|
449 |
'prf' options thmrefs |
|
berghofe@13827
|
450 |
'full\_prf' options thmrefs
|
wenzelm@9200
|
451 |
;
|
wenzelm@9200
|
452 |
options: '[' (option * ',') ']'
|
wenzelm@9200
|
453 |
;
|
wenzelm@9200
|
454 |
option: name | name '=' name
|
wenzelm@9200
|
455 |
;
|
wenzelm@9200
|
456 |
\end{rail}
|
wenzelm@9200
|
457 |
|
wenzelm@9200
|
458 |
Note that the syntax of antiquotations may \emph{not} include source comments
|
wenzelm@9200
|
459 |
\texttt{(*~\dots~*)} or verbatim text \verb|{*|~\dots~\verb|*}|.
|
wenzelm@9200
|
460 |
|
wenzelm@10319
|
461 |
\begin{descr}
|
wenzelm@14895
|
462 |
|
wenzelm@10336
|
463 |
\item [$\at\{thm~\vec a\}$] prints theorems $\vec a$. Note that attribute
|
wenzelm@10336
|
464 |
specifications may be included as well (see also \S\ref{sec:syn-att}); the
|
wenzelm@12618
|
465 |
$no_vars$ operation (see \S\ref{sec:misc-meth-att}) would be particularly
|
wenzelm@10336
|
466 |
useful to suppress printing of schematic variables.
|
wenzelm@13039
|
467 |
|
wenzelm@10336
|
468 |
\item [$\at\{prop~\phi\}$] prints a well-typed proposition $\phi$.
|
wenzelm@13039
|
469 |
|
wenzelm@10336
|
470 |
\item [$\at\{term~t\}$] prints a well-typed term $t$.
|
wenzelm@13039
|
471 |
|
wenzelm@10336
|
472 |
\item [$\at\{typ~\tau\}$] prints a well-formed type $\tau$.
|
wenzelm@14895
|
473 |
|
wenzelm@10336
|
474 |
\item [$\at\{text~s\}$] prints uninterpreted source text $s$. This is
|
wenzelm@10336
|
475 |
particularly useful to print portions of text according to the Isabelle
|
wenzelm@10336
|
476 |
{\LaTeX} output style, without demanding well-formedness (e.g.\ small pieces
|
wenzelm@13039
|
477 |
of terms that should not be parsed or type-checked yet).
|
wenzelm@14895
|
478 |
|
wenzelm@10336
|
479 |
\item [$\at\{goals\}$] prints the current \emph{dynamic} goal state. This is
|
wenzelm@13039
|
480 |
mainly for support of tactic-emulation scripts within Isar --- presentation
|
wenzelm@13039
|
481 |
of goal states does not conform to actual human-readable proof documents.
|
wenzelm@10319
|
482 |
Please do not include goal states into document output unless you really
|
wenzelm@10319
|
483 |
know what you are doing!
|
wenzelm@13039
|
484 |
|
wenzelm@10355
|
485 |
\item [$\at\{subgoals\}$] behaves almost like $goals$, except that it does not
|
wenzelm@10355
|
486 |
print the main goal.
|
wenzelm@13039
|
487 |
|
berghofe@13827
|
488 |
\item [$\at\{prf~\vec a\}$] prints the (compact) proof terms corresponding to
|
berghofe@13827
|
489 |
the theorems $\vec a$. Note that this
|
berghofe@13827
|
490 |
requires proof terms to be switched on for the current object logic
|
berghofe@13827
|
491 |
(see the ``Proof terms'' section of the Isabelle reference manual
|
berghofe@13827
|
492 |
for information on how to do this).
|
berghofe@13827
|
493 |
|
berghofe@13827
|
494 |
\item [$\at\{full_prf~\vec a\}$] is like $\at\{prf~\vec a\}$, but displays
|
berghofe@13827
|
495 |
the full proof terms, i.e.\ also displays information omitted in
|
berghofe@13827
|
496 |
the compact proof term, which is denoted by ``$_$'' placeholders there.
|
berghofe@13827
|
497 |
|
wenzelm@10319
|
498 |
\end{descr}
|
wenzelm@10319
|
499 |
|
wenzelm@9200
|
500 |
\medskip
|
wenzelm@9200
|
501 |
|
wenzelm@10336
|
502 |
The following options are available to tune the output. Note that most of
|
wenzelm@9233
|
503 |
these coincide with ML flags of the same names (see also \cite{isabelle-ref}).
|
wenzelm@9200
|
504 |
\begin{descr}
|
wenzelm@9233
|
505 |
\item[$show_types = bool$ and $show_sorts = bool$] control printing of
|
wenzelm@9234
|
506 |
explicit type and sort constraints.
|
wenzelm@14707
|
507 |
\item[$show_structs = bool$] controls printing of implicit structures.
|
wenzelm@9233
|
508 |
\item[$long_names = bool$] forces names of types and constants etc.\ to be
|
wenzelm@9233
|
509 |
printed in their fully qualified internal form.
|
wenzelm@9233
|
510 |
\item[$eta_contract = bool$] prints terms in $\eta$-contracted form.
|
wenzelm@9200
|
511 |
\item[$display = bool$] indicates if the text is to be output as multi-line
|
wenzelm@9200
|
512 |
``display material'', rather than a small piece of text without line breaks
|
wenzelm@9200
|
513 |
(which is the default).
|
wenzelm@14689
|
514 |
\item[$breaks = bool$] controls line breaks in non-display material.
|
wenzelm@9200
|
515 |
\item[$quotes = bool$] indicates if the output should be enclosed in double
|
wenzelm@9200
|
516 |
quotes.
|
wenzelm@9233
|
517 |
\item[$mode = name$] adds $name$ to the print mode to be used for presentation
|
wenzelm@9233
|
518 |
(see also \cite{isabelle-ref}). Note that the standard setup for {\LaTeX}
|
wenzelm@9233
|
519 |
output is already present by default, including the modes ``$latex$'',
|
wenzelm@9233
|
520 |
``$xsymbols$'', ``$symbols$''.
|
wenzelm@9728
|
521 |
\item[$margin = nat$ and $indent = nat$] change the margin or indentation for
|
wenzelm@9728
|
522 |
pretty printing of display material.
|
wenzelm@9752
|
523 |
\item[$source = bool$] prints the source text of the antiquotation arguments,
|
wenzelm@9752
|
524 |
rather than the actual value. Note that this does not affect
|
wenzelm@9752
|
525 |
well-formedness checks of $thm$, $term$, etc. (only the $text$ antiquotation
|
wenzelm@9752
|
526 |
admits arbitrary output).
|
wenzelm@10319
|
527 |
\item[$goals_limit = nat$] determines the maximum number of goals to be
|
wenzelm@10319
|
528 |
printed.
|
wenzelm@14895
|
529 |
\item[$locale = name$] specifies an alternative context used for evaluating
|
wenzelm@14919
|
530 |
and printing the subsequent argument.
|
wenzelm@9200
|
531 |
\end{descr}
|
wenzelm@9200
|
532 |
|
wenzelm@9200
|
533 |
For boolean flags, ``$name = true$'' may be abbreviated as ``$name$''. All of
|
wenzelm@9200
|
534 |
the above flags are disabled by default, unless changed from ML.
|
wenzelm@9200
|
535 |
|
wenzelm@10336
|
536 |
\medskip Note that antiquotations do not only spare the author from tedious
|
wenzelm@13039
|
537 |
typing of logical entities, but also achieve some degree of
|
wenzelm@13039
|
538 |
consistency-checking of informal explanations with formal developments:
|
wenzelm@13039
|
539 |
well-formedness of terms and types with respect to the current theory or proof
|
wenzelm@13039
|
540 |
context is ensured here.
|
wenzelm@9200
|
541 |
|
wenzelm@14895
|
542 |
%%% Local Variables:
|
wenzelm@7046
|
543 |
%%% mode: latex
|
wenzelm@7046
|
544 |
%%% TeX-master: "isar-ref"
|
wenzelm@14895
|
545 |
%%% End:
|