wenzelm@7046
|
1 |
|
wenzelm@7167
|
2 |
\chapter{Basic Isar elements}
|
wenzelm@7046
|
3 |
|
wenzelm@7167
|
4 |
Subsequently, we introduce most of the basic Isar theory and proof commands as
|
wenzelm@7167
|
5 |
provided by Isabelle/Pure. Chapter~\ref{ch:gen-tools} describes further Isar
|
wenzelm@7167
|
6 |
elements as provided by generic tools and packages that are either part of
|
wenzelm@7167
|
7 |
Pure Isabelle, or preloaded by most object logics (such as the simplifier).
|
wenzelm@7167
|
8 |
See chapter~\ref{ch:hol-tools} for actual object-logic specific elements (for
|
wenzelm@7167
|
9 |
Isabelle/HOL).
|
wenzelm@7167
|
10 |
|
wenzelm@7167
|
11 |
\medskip
|
wenzelm@7167
|
12 |
|
wenzelm@7167
|
13 |
Isar commands may be either \emph{proper} document constructors, or
|
wenzelm@7167
|
14 |
\emph{improper commands} (indicated by $^*$). Improper commands might be
|
wenzelm@7167
|
15 |
helpful when developing proof documents, while their use is strongly
|
wenzelm@7167
|
16 |
discouraged for the final outcome. Typical examples are diagnostic commands
|
wenzelm@7167
|
17 |
that print terms or theorems according to the current context; other commands
|
wenzelm@7167
|
18 |
even emulate old-style tactical theorem proving, which facilitates porting of
|
wenzelm@7167
|
19 |
legacy proof scripts.
|
wenzelm@7167
|
20 |
|
wenzelm@7134
|
21 |
|
wenzelm@7134
|
22 |
\section{Theory commands}
|
wenzelm@7134
|
23 |
|
wenzelm@7167
|
24 |
\subsection{Defining theories}\label{sec:begin-thy}
|
wenzelm@7134
|
25 |
|
wenzelm@7134
|
26 |
\indexisarcmd{theory}\indexisarcmd{end}\indexisarcmd{context}
|
wenzelm@7134
|
27 |
\begin{matharray}{rcl}
|
wenzelm@7134
|
28 |
\isarcmd{theory} & : & \isartrans{\cdot}{theory} \\
|
wenzelm@7134
|
29 |
\isarcmd{context}^* & : & \isartrans{\cdot}{theory} \\
|
wenzelm@7134
|
30 |
\isarcmd{end} & : & \isartrans{theory}{\cdot} \\
|
wenzelm@7134
|
31 |
\end{matharray}
|
wenzelm@7134
|
32 |
|
wenzelm@7134
|
33 |
Isabelle/Isar ``new-style'' theories are either defined via theory files or
|
wenzelm@7167
|
34 |
interactively. Both actual theory specifications and proofs are handled
|
wenzelm@7167
|
35 |
uniformly --- occasionally definitional mechanisms even require some proof.
|
wenzelm@7167
|
36 |
In contrast, ``old-style'' Isabelle theories support batch processing only,
|
wenzelm@7167
|
37 |
with the proof scripts collected in separate ML files.
|
wenzelm@7134
|
38 |
|
wenzelm@7134
|
39 |
The first command of any theory has to be $\THEORY$, starting a new theory
|
wenzelm@7167
|
40 |
based on the merge of existing ones. The theory context may be changed by
|
wenzelm@7167
|
41 |
$\CONTEXT$ without creating a new theory. In both cases $\END$ concludes the
|
wenzelm@7167
|
42 |
theory development; it has to be the very last command of any proper theory
|
wenzelm@7167
|
43 |
file.
|
wenzelm@7134
|
44 |
|
wenzelm@7134
|
45 |
\begin{rail}
|
wenzelm@7134
|
46 |
'theory' name '=' (name + '+') filespecs? ':'
|
wenzelm@7134
|
47 |
;
|
wenzelm@7134
|
48 |
'context' name
|
wenzelm@7134
|
49 |
;
|
wenzelm@7134
|
50 |
'end'
|
wenzelm@7134
|
51 |
;;
|
wenzelm@7134
|
52 |
|
wenzelm@7167
|
53 |
filespecs: 'files' ((name | parname) +);
|
wenzelm@7134
|
54 |
\end{rail}
|
wenzelm@7134
|
55 |
|
wenzelm@7167
|
56 |
\begin{descr}
|
wenzelm@7134
|
57 |
\item [$\THEORY~A = B@1 + \cdots + B@n$] commences a new theory $A$ based on
|
wenzelm@7134
|
58 |
existing ones $B@1 + \cdots + B@n$. Note that Isabelle's theory loader
|
wenzelm@7134
|
59 |
system ensures that any of the base theories are properly loaded (and fully
|
wenzelm@7167
|
60 |
up-to-date when $\THEORY$ is executed interactively). The optional
|
wenzelm@7167
|
61 |
$\isarkeyword{files}$ specification declares additional dependencies on ML
|
wenzelm@7167
|
62 |
files. Unless put in in parentheses, any file will be loaded immediately
|
wenzelm@7167
|
63 |
via $\isarcmd{use}$ (see also \S\ref{sec:ML}).
|
wenzelm@7134
|
64 |
|
wenzelm@7167
|
65 |
\item [$\CONTEXT~B$] enters an existing theory context $B$, basically in
|
wenzelm@7134
|
66 |
read-only mode, so only a limited set of commands may be performed. Just as
|
wenzelm@7134
|
67 |
for $\THEORY$, the theory loader ensures that $B$ is loaded and up-to-date.
|
wenzelm@7134
|
68 |
|
wenzelm@7167
|
69 |
\item [$\END$] concludes the current theory definition or context switch.
|
wenzelm@7167
|
70 |
\end{descr}
|
wenzelm@7134
|
71 |
|
wenzelm@7134
|
72 |
|
wenzelm@7167
|
73 |
\subsection{Formal comments}\label{sec:formal-cmt-thy}
|
wenzelm@7134
|
74 |
|
wenzelm@7167
|
75 |
\indexisarcmd{title}\indexisarcmd{chapter}\indexisarcmd{section}\indexisarcmd{subsection}
|
wenzelm@7167
|
76 |
\indexisarcmd{subsubsection}\indexisarcmd{text}
|
wenzelm@7134
|
77 |
\begin{matharray}{rcl}
|
wenzelm@7134
|
78 |
\isarcmd{title} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
79 |
\isarcmd{chapter} & : & \isartrans{theory}{theory} \\
|
wenzelm@7167
|
80 |
\isarcmd{section} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
81 |
\isarcmd{subsection} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
82 |
\isarcmd{subsubsection} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
83 |
\isarcmd{text} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
84 |
\end{matharray}
|
wenzelm@7134
|
85 |
|
wenzelm@7167
|
86 |
There are several commands to include \emph{formal comments} in theory
|
wenzelm@7167
|
87 |
specification (a few more are available for proofs, see
|
wenzelm@7167
|
88 |
\S\ref{sec:formal-cmt-prf}). In contrast to source-level comments
|
wenzelm@7134
|
89 |
\verb|(*|\dots\verb|*)|, which are stripped at the lexical level, any text
|
wenzelm@7134
|
90 |
given as formal comment is meant to be part of the actual document.
|
wenzelm@7134
|
91 |
Consequently, it would be included in the final printed version.
|
wenzelm@7134
|
92 |
|
wenzelm@7134
|
93 |
Apart from plain prose, formal comments may also refer to logical entities of
|
wenzelm@7134
|
94 |
the current theory context (types, terms, theorems etc.). Proper processing
|
wenzelm@7134
|
95 |
of the text would then include some further consistency checks with the items
|
wenzelm@7134
|
96 |
declared in the current theory, e.g.\ type-checking of included terms.
|
wenzelm@7134
|
97 |
\footnote{The current version of Isabelle/Isar does not process formal
|
wenzelm@7134
|
98 |
comments in any such way. This will be available as part of the automatic
|
wenzelm@7134
|
99 |
theory and proof document preparation system (via (PDF)LaTeX) that is
|
wenzelm@7134
|
100 |
planned for the near future.}
|
wenzelm@7134
|
101 |
|
wenzelm@7134
|
102 |
\begin{rail}
|
wenzelm@7134
|
103 |
'title' text text? text?
|
wenzelm@7134
|
104 |
;
|
wenzelm@7167
|
105 |
('chapter' | 'section' | 'subsection' | 'subsubsection' | 'text') text
|
wenzelm@7134
|
106 |
;
|
wenzelm@7134
|
107 |
\end{rail}
|
wenzelm@7134
|
108 |
|
wenzelm@7167
|
109 |
\begin{descr}
|
wenzelm@7134
|
110 |
\item [$\isarkeyword{title}~title~author~date$] specifies the document title
|
wenzelm@7134
|
111 |
just as in typical LaTeX documents.
|
wenzelm@7167
|
112 |
\item [$\isarkeyword{chapter}~text$, $\isarkeyword{section}~text$,
|
wenzelm@7167
|
113 |
$\isarkeyword{subsection}~text$, $\isarkeyword{subsubsection}~text$] specify
|
wenzelm@7167
|
114 |
chapter and subsection headings.
|
wenzelm@7134
|
115 |
\item [$\TEXT~text$] specifies an actual body of prose text, including
|
wenzelm@7134
|
116 |
references to formal entities.\footnote{The latter feature is not yet
|
wenzelm@7134
|
117 |
exploited in any way.}
|
wenzelm@7167
|
118 |
\end{descr}
|
wenzelm@7134
|
119 |
|
wenzelm@7134
|
120 |
|
wenzelm@7135
|
121 |
\subsection{Type classes and sorts}\label{sec:classes}
|
wenzelm@7134
|
122 |
|
wenzelm@7134
|
123 |
\indexisarcmd{classes}\indexisarcmd{classrel}\indexisarcmd{defaultsort}
|
wenzelm@7134
|
124 |
\begin{matharray}{rcl}
|
wenzelm@7134
|
125 |
\isarcmd{classes} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
126 |
\isarcmd{classrel} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
127 |
\isarcmd{defaultsort} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
128 |
\end{matharray}
|
wenzelm@7134
|
129 |
|
wenzelm@7134
|
130 |
\begin{rail}
|
wenzelm@7167
|
131 |
'classes' (classdecl comment? +)
|
wenzelm@7134
|
132 |
;
|
wenzelm@7134
|
133 |
'classrel' nameref '<' nameref comment?
|
wenzelm@7134
|
134 |
;
|
wenzelm@7134
|
135 |
'defaultsort' sort comment?
|
wenzelm@7134
|
136 |
;
|
wenzelm@7134
|
137 |
\end{rail}
|
wenzelm@7134
|
138 |
|
wenzelm@7167
|
139 |
\begin{descr}
|
wenzelm@7134
|
140 |
\item [$\isarkeyword{classes}~c<cs ~\dots$] declares class $c$ to be a
|
wenzelm@7134
|
141 |
subclass of existing classes $cs$. Cyclic class structures are ruled out.
|
wenzelm@7134
|
142 |
\item [$\isarkeyword{classrel}~c@1<c@2$] states a subclass relation between
|
wenzelm@7134
|
143 |
existing classes $c@1$ and $c@2$. This is done axiomatically! The
|
wenzelm@7134
|
144 |
$\isarkeyword{instance}$ command provides a way introduce proven class
|
wenzelm@7134
|
145 |
relations (see \S\ref{sec:axclass}).
|
wenzelm@7134
|
146 |
\item [$\isarkeyword{defaultsort}~s$] makes sort $s$ the new default sort for
|
wenzelm@7134
|
147 |
any type variables input without sort constraints. Typically, the default
|
wenzelm@7134
|
148 |
sort would be only changed when defining new logics.
|
wenzelm@7167
|
149 |
\end{descr}
|
wenzelm@7134
|
150 |
|
wenzelm@7134
|
151 |
|
wenzelm@7141
|
152 |
\subsection{Types}\label{sec:types-pure}
|
wenzelm@7134
|
153 |
|
wenzelm@7134
|
154 |
\indexisarcmd{typedecl}\indexisarcmd{types}\indexisarcmd{nonterminals}\indexisarcmd{arities}
|
wenzelm@7134
|
155 |
\begin{matharray}{rcl}
|
wenzelm@7134
|
156 |
\isarcmd{types} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
157 |
\isarcmd{typedecl} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
158 |
\isarcmd{nonterminals} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
159 |
\isarcmd{arities} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
160 |
\end{matharray}
|
wenzelm@7134
|
161 |
|
wenzelm@7134
|
162 |
\begin{rail}
|
wenzelm@7134
|
163 |
'types' (typespec '=' type infix? comment? +)
|
wenzelm@7134
|
164 |
;
|
wenzelm@7134
|
165 |
'typedecl' typespec infix? comment?
|
wenzelm@7134
|
166 |
;
|
wenzelm@7134
|
167 |
'nonterminals' (name +) comment?
|
wenzelm@7134
|
168 |
;
|
wenzelm@7134
|
169 |
'arities' (nameref '::' arity comment? +)
|
wenzelm@7134
|
170 |
;
|
wenzelm@7134
|
171 |
\end{rail}
|
wenzelm@7134
|
172 |
|
wenzelm@7167
|
173 |
\begin{descr}
|
wenzelm@7134
|
174 |
\item [$\TYPES~(\vec\alpha)t = \tau~\dots$] introduces \emph{type synonym}
|
wenzelm@7134
|
175 |
$(\vec\alpha)t$ for existing type $\tau$. Unlike actual type definitions,
|
wenzelm@7134
|
176 |
as are available in Isabelle/HOL for example, type synonyms are just purely
|
wenzelm@7134
|
177 |
syntactic abbreviations, without any logical significance. Internally, type
|
wenzelm@7134
|
178 |
synonyms are fully expanded, as may be observed when printing terms or
|
wenzelm@7134
|
179 |
theorems.
|
wenzelm@7134
|
180 |
\item [$\isarkeyword{typedecl}~(\vec\alpha)t$] declares a new type constructor
|
wenzelm@7134
|
181 |
$t$, intended as an actual logical type. Note that some logics such as
|
wenzelm@7134
|
182 |
Isabelle/HOL provide their own version of $\isarkeyword{typedecl}$.
|
wenzelm@7134
|
183 |
\item [$\isarkeyword{nonterminals}~c~\dots$] declares $0$-ary type
|
wenzelm@7134
|
184 |
constructors $c$ to act as purely syntactic types, i.e.\ nonterminal symbols
|
wenzelm@7134
|
185 |
of Isabelle's inner syntax of terms or types.
|
wenzelm@7134
|
186 |
\item [$\isarkeyword{arities}~t::(\vec s)s~\dots$] augments Isabelle's
|
wenzelm@7134
|
187 |
order-sorted signature of types by new type constructor arities. This is
|
wenzelm@7134
|
188 |
done axiomatically! The $\isarkeyword{instance}$ command provides a way
|
wenzelm@7134
|
189 |
introduce proven type arities (see \S\ref{sec:axclass}).
|
wenzelm@7167
|
190 |
\end{descr}
|
wenzelm@7134
|
191 |
|
wenzelm@7134
|
192 |
|
wenzelm@7134
|
193 |
\subsection{Constants and simple definitions}
|
wenzelm@7134
|
194 |
|
wenzelm@7134
|
195 |
\indexisarcmd{consts}\indexisarcmd{defs}\indexisarcmd{constdefs}
|
wenzelm@7134
|
196 |
\begin{matharray}{rcl}
|
wenzelm@7134
|
197 |
\isarcmd{consts} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
198 |
\isarcmd{defs} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
199 |
\isarcmd{constdefs} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
200 |
\end{matharray}
|
wenzelm@7134
|
201 |
|
wenzelm@7134
|
202 |
\begin{rail}
|
wenzelm@7134
|
203 |
'consts' (constdecl +)
|
wenzelm@7134
|
204 |
;
|
wenzelm@7134
|
205 |
'defs' (thmdecl? prop comment? +)
|
wenzelm@7134
|
206 |
;
|
wenzelm@7134
|
207 |
'constdefs' (constdecl prop comment? +)
|
wenzelm@7134
|
208 |
;
|
wenzelm@7134
|
209 |
|
wenzelm@7134
|
210 |
constdecl: name '::' type mixfix? comment?
|
wenzelm@7134
|
211 |
;
|
wenzelm@7134
|
212 |
\end{rail}
|
wenzelm@7134
|
213 |
|
wenzelm@7167
|
214 |
\begin{descr}
|
wenzelm@7134
|
215 |
\item [$\CONSTS~c::\tau~\dots$] declares constant $c$ to have any instance of
|
wenzelm@7134
|
216 |
type scheme $\tau$. The optional mixfix annotations may attach concrete
|
wenzelm@7134
|
217 |
syntax to the constant.
|
wenzelm@7134
|
218 |
\item [$\DEFS~name: eqn~\dots$] introduces $eqn$ as a definitional axiom for
|
wenzelm@7134
|
219 |
some existing constant. See \cite[\S6]{isabelle-ref} for more details on
|
wenzelm@7134
|
220 |
the form of equations admitted as constant definitions.
|
wenzelm@7134
|
221 |
\item [$\isarkeyword{constdefs}~c::\tau~eqn~\dots$] combines constant
|
wenzelm@7134
|
222 |
declarations and definitions, using canonical name $c_def$ for the
|
wenzelm@7134
|
223 |
definitional axiom.
|
wenzelm@7167
|
224 |
\end{descr}
|
wenzelm@7134
|
225 |
|
wenzelm@7134
|
226 |
|
wenzelm@7167
|
227 |
\subsection{Syntax and translations}
|
wenzelm@7134
|
228 |
|
wenzelm@7134
|
229 |
\indexisarcmd{syntax}\indexisarcmd{translations}
|
wenzelm@7134
|
230 |
\begin{matharray}{rcl}
|
wenzelm@7134
|
231 |
\isarcmd{syntax} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
232 |
\isarcmd{translations} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
233 |
\end{matharray}
|
wenzelm@7134
|
234 |
|
wenzelm@7134
|
235 |
\begin{rail}
|
wenzelm@7134
|
236 |
'syntax' ('(' name 'output'? ')')? (constdecl +)
|
wenzelm@7134
|
237 |
;
|
wenzelm@7134
|
238 |
'translations' (transpat ('==' | '=>' | '<=') transpat comment? +)
|
wenzelm@7134
|
239 |
;
|
wenzelm@7134
|
240 |
transpat: ('(' nameref ')')? string
|
wenzelm@7134
|
241 |
;
|
wenzelm@7134
|
242 |
\end{rail}
|
wenzelm@7134
|
243 |
|
wenzelm@7167
|
244 |
\begin{descr}
|
wenzelm@7134
|
245 |
\item [$\isarkeyword{syntax}~mode~decls$] is similar to $\CONSTS~decls$,
|
wenzelm@7134
|
246 |
except the actual logical signature extension. Thus the context free
|
wenzelm@7134
|
247 |
grammar of Isabelle's inner syntax may be augmented in arbitrary ways. The
|
wenzelm@7134
|
248 |
$mode$ argument refers to the print mode that the grammar rules belong;
|
wenzelm@7134
|
249 |
unless there is the \texttt{output} flag given, all productions are added
|
wenzelm@7134
|
250 |
both to the input and output grammar.
|
wenzelm@7134
|
251 |
\item [$\isarkeyword{translations}~rule~\dots$] specifies syntactic
|
wenzelm@7134
|
252 |
translation rules (macros): parse/print rules (\texttt{==}), parse rules
|
wenzelm@7134
|
253 |
(\texttt{=>}), print rules (\texttt{<=}). Translation patterns may be
|
wenzelm@7134
|
254 |
prefixed by the syntactic category to be used for parsing; the default is
|
wenzelm@7134
|
255 |
\texttt{logic}.
|
wenzelm@7167
|
256 |
\end{descr}
|
wenzelm@7134
|
257 |
|
wenzelm@7134
|
258 |
|
wenzelm@7134
|
259 |
\subsection{Axioms and theorems}
|
wenzelm@7134
|
260 |
|
wenzelm@7134
|
261 |
\indexisarcmd{axioms}\indexisarcmd{theorems}\indexisarcmd{lemmas}
|
wenzelm@7134
|
262 |
\begin{matharray}{rcl}
|
wenzelm@7134
|
263 |
\isarcmd{axioms} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
264 |
\isarcmd{theorems} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
265 |
\isarcmd{lemmas} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
266 |
\end{matharray}
|
wenzelm@7134
|
267 |
|
wenzelm@7134
|
268 |
\begin{rail}
|
wenzelm@7135
|
269 |
'axioms' (axmdecl prop comment? +)
|
wenzelm@7134
|
270 |
;
|
wenzelm@7134
|
271 |
('theorems' | 'lemmas') thmdef? thmrefs
|
wenzelm@7134
|
272 |
;
|
wenzelm@7134
|
273 |
\end{rail}
|
wenzelm@7134
|
274 |
|
wenzelm@7167
|
275 |
\begin{descr}
|
wenzelm@7134
|
276 |
\item [$\isarkeyword{axioms}~name: \phi~\dots$] introduces arbitrary
|
wenzelm@7134
|
277 |
statements as logical axioms. In fact, axioms are ``axiomatic theorems'',
|
wenzelm@7134
|
278 |
and may be referred as any other theorems later.
|
wenzelm@7134
|
279 |
|
wenzelm@7134
|
280 |
Axioms are usually only introduced when declaring new logical systems.
|
wenzelm@7134
|
281 |
Everyday work is normally done the hard way, with proper definitions and
|
wenzelm@7134
|
282 |
actual theorems.
|
wenzelm@7134
|
283 |
\item [$\isarkeyword{theorems}~name = thms$] stores lists of existing theorems
|
wenzelm@7134
|
284 |
as $name$. Typical applications would also involve attributes to augment
|
wenzelm@7134
|
285 |
the default simpset, for example.
|
wenzelm@7134
|
286 |
\item [$\isarkeyword{lemmas}$] is similar to $\isarkeyword{theorems}$, but
|
wenzelm@7134
|
287 |
tags the results as ``lemma''.
|
wenzelm@7167
|
288 |
\end{descr}
|
wenzelm@7134
|
289 |
|
wenzelm@7134
|
290 |
|
wenzelm@7167
|
291 |
\subsection{Name spaces}
|
wenzelm@7134
|
292 |
|
wenzelm@7167
|
293 |
Isabelle organizes any kind of names (of types, constants, theorems etc.) by
|
wenzelm@7167
|
294 |
hierarchically structured name spaces. Normally the user never has to control
|
wenzelm@7167
|
295 |
the behavior of name space entry by hand, yet the following commands provide
|
wenzelm@7167
|
296 |
some way to do so.
|
wenzelm@7167
|
297 |
|
wenzelm@7167
|
298 |
\indexisarcmd{global}\indexisarcmd{local}
|
wenzelm@7134
|
299 |
\begin{matharray}{rcl}
|
wenzelm@7134
|
300 |
\isarcmd{global} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
301 |
\isarcmd{local} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
302 |
\end{matharray}
|
wenzelm@7134
|
303 |
|
wenzelm@7167
|
304 |
\begin{descr}
|
wenzelm@7167
|
305 |
\item [$\isarkeyword{global}$ and $\isarkeyword{local}$] change the current
|
wenzelm@7167
|
306 |
name declaration mode. Initially, theories start in $\isarkeyword{local}$
|
wenzelm@7167
|
307 |
mode, causing all names to be automatically qualified by the theory name.
|
wenzelm@7167
|
308 |
Changing this to $\isarkeyword{global}$ causes all names to be declared as
|
wenzelm@7167
|
309 |
base names only.
|
wenzelm@7167
|
310 |
\end{descr}
|
wenzelm@7134
|
311 |
|
wenzelm@7134
|
312 |
|
wenzelm@7167
|
313 |
\subsection{Incorporating ML code}\label{sec:ML}
|
wenzelm@7134
|
314 |
|
wenzelm@7134
|
315 |
\indexisarcmd{use}\indexisarcmd{ML}\indexisarcmd{setup}
|
wenzelm@7134
|
316 |
\begin{matharray}{rcl}
|
wenzelm@7134
|
317 |
\isarcmd{use} & : & \isartrans{\cdot}{\cdot} \\
|
wenzelm@7134
|
318 |
\isarcmd{ML} & : & \isartrans{\cdot}{\cdot} \\
|
wenzelm@7134
|
319 |
\isarcmd{setup} & : & \isartrans{\cdot}{\cdot} \\
|
wenzelm@7134
|
320 |
\end{matharray}
|
wenzelm@7134
|
321 |
|
wenzelm@7134
|
322 |
\begin{rail}
|
wenzelm@7134
|
323 |
'use' name
|
wenzelm@7134
|
324 |
;
|
wenzelm@7134
|
325 |
'ML' text
|
wenzelm@7134
|
326 |
;
|
wenzelm@7134
|
327 |
'setup' text
|
wenzelm@7134
|
328 |
;
|
wenzelm@7134
|
329 |
\end{rail}
|
wenzelm@7134
|
330 |
|
wenzelm@7167
|
331 |
\begin{descr}
|
wenzelm@7167
|
332 |
\item [$\isarkeyword{use}~file$] reads and execute ML commands from $file$.
|
wenzelm@7167
|
333 |
The current theory context as passed down to the ML session. Furthermore,
|
wenzelm@7167
|
334 |
the file name is checked with the dependency declarations given in the
|
wenzelm@7167
|
335 |
theory header (see also \S\ref{sec:begin-thy}).
|
wenzelm@7167
|
336 |
\item [$\isarkeyword{ML}~text$] reads and executes ML commands from $text$.
|
wenzelm@7167
|
337 |
The theory context is passed just as in $\isarkeyword{use}$.
|
wenzelm@7167
|
338 |
\item [$\isarkeyword{setup}~text$] changes the current theory context by
|
wenzelm@7167
|
339 |
applying setup functions $text$ (which has to be an ML expression of type
|
wenzelm@7167
|
340 |
$(theory -> theory)~list$. The $\isarkeyword{setup}$ is the usual way to
|
wenzelm@7167
|
341 |
initialize object-logic specific tools and packages written in ML.
|
wenzelm@7167
|
342 |
\end{descr}
|
wenzelm@7134
|
343 |
|
wenzelm@7134
|
344 |
|
wenzelm@7167
|
345 |
\subsection{Syntax translation functions}
|
wenzelm@7134
|
346 |
|
wenzelm@7167
|
347 |
\indexisarcmd{parse-ast-translation}\indexisarcmd{parse-translation}
|
wenzelm@7167
|
348 |
\indexisarcmd{print-translation}\indexisarcmd{typed-print-translation}
|
wenzelm@7167
|
349 |
\indexisarcmd{print-ast-translation}\indexisarcmd{token-translation}
|
wenzelm@7134
|
350 |
\begin{matharray}{rcl}
|
wenzelm@7134
|
351 |
\isarcmd{parse_ast_translation} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
352 |
\isarcmd{parse_translation} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
353 |
\isarcmd{print_translation} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
354 |
\isarcmd{typed_print_translation} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
355 |
\isarcmd{print_ast_translation} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
356 |
\isarcmd{token_translation} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
357 |
\end{matharray}
|
wenzelm@7134
|
358 |
|
wenzelm@7134
|
359 |
Syntax translation functions written in ML admit almost arbitrary
|
wenzelm@7134
|
360 |
manipulations of Isabelle's inner syntax. Any of the above commands have a
|
wenzelm@7134
|
361 |
single \railqtoken{text} argument that refers to an ML expression of
|
wenzelm@7134
|
362 |
appropriate type. See \cite[\S8]{isabelle-ref} for more information on syntax
|
wenzelm@7134
|
363 |
transformations.
|
wenzelm@7134
|
364 |
|
wenzelm@7134
|
365 |
|
wenzelm@7134
|
366 |
\subsection{Oracles}
|
wenzelm@7134
|
367 |
|
wenzelm@7134
|
368 |
\indexisarcmd{oracle}
|
wenzelm@7134
|
369 |
\begin{matharray}{rcl}
|
wenzelm@7134
|
370 |
\isarcmd{oracle} & : & \isartrans{theory}{theory} \\
|
wenzelm@7134
|
371 |
\end{matharray}
|
wenzelm@7134
|
372 |
|
wenzelm@7134
|
373 |
\begin{rail}
|
wenzelm@7134
|
374 |
'oracle' name '=' text comment?
|
wenzelm@7134
|
375 |
;
|
wenzelm@7134
|
376 |
\end{rail}
|
wenzelm@7134
|
377 |
|
wenzelm@7167
|
378 |
\begin{descr}
|
wenzelm@7134
|
379 |
\item [$\isarkeyword{oracle}~name=text$] FIXME
|
wenzelm@7167
|
380 |
\end{descr}
|
wenzelm@7134
|
381 |
|
wenzelm@7134
|
382 |
|
wenzelm@7134
|
383 |
\section{Proof commands}
|
wenzelm@7134
|
384 |
|
wenzelm@7167
|
385 |
Proof commands provide transitions of Isar/VM machine configurations. There
|
wenzelm@7167
|
386 |
are three different kinds of operation:
|
wenzelm@7167
|
387 |
\begin{descr}
|
wenzelm@7167
|
388 |
\item [$proof(prove)$] means that a new goal has just been stated that is now
|
wenzelm@7167
|
389 |
to be \emph{proven}; the next command may refine it by some proof method
|
wenzelm@7167
|
390 |
($\approx$ tactic) and enter a sub-proof to establish the final result.
|
wenzelm@7167
|
391 |
\item [$proof(state)$] is like an internal theory mode: the context may be
|
wenzelm@7167
|
392 |
augmented by \emph{stating} additional assumptions, intermediate result;
|
wenzelm@7167
|
393 |
\item [$proof(chain)$] indicates an intermediate mode between $proof(state)$
|
wenzelm@7167
|
394 |
and $proof(state)$: some already established facts have been just picked up
|
wenzelm@7167
|
395 |
in order to use them when refining the subsequent goal.
|
wenzelm@7167
|
396 |
\end{descr}
|
wenzelm@7134
|
397 |
|
wenzelm@7167
|
398 |
|
wenzelm@7167
|
399 |
\subsection{Formal comments}\label{sec:formal-cmt-prf}
|
wenzelm@7167
|
400 |
|
wenzelm@7167
|
401 |
The following formal comments in proof mode closely correspond to the ones of
|
wenzelm@7167
|
402 |
theory mode (see \S\ref{sec:formal-cmt-thy} for more information).
|
wenzelm@7167
|
403 |
|
wenzelm@7167
|
404 |
\indexisarcmd{sect}\indexisarcmd{subsect}\indexisarcmd{subsect}\indexisarcmd{txt}
|
wenzelm@7134
|
405 |
\begin{matharray}{rcl}
|
wenzelm@7167
|
406 |
\isarcmd{sect} & : & \isartrans{proof(state)}{proof(state)} \\
|
wenzelm@7167
|
407 |
\isarcmd{subsect} & : & \isartrans{proof(state)}{proof(state)} \\
|
wenzelm@7167
|
408 |
\isarcmd{subsubsect} & : & \isartrans{proof(state)}{proof(state)} \\
|
wenzelm@7167
|
409 |
\isarcmd{txt} & : & \isartrans{proof(state)}{proof(state)} \\
|
wenzelm@7134
|
410 |
\end{matharray}
|
wenzelm@7134
|
411 |
|
wenzelm@7134
|
412 |
\begin{rail}
|
wenzelm@7167
|
413 |
('sect' | 'subsect' | 'subsubsect' | 'txt') text
|
wenzelm@7134
|
414 |
;
|
wenzelm@7134
|
415 |
\end{rail}
|
wenzelm@7134
|
416 |
|
wenzelm@7134
|
417 |
|
wenzelm@7134
|
418 |
\subsection{Proof context}
|
wenzelm@7134
|
419 |
|
wenzelm@7167
|
420 |
FIXME
|
wenzelm@7167
|
421 |
|
wenzelm@7134
|
422 |
\indexisarcmd{fix}\indexisarcmd{assume}\indexisarcmd{presume}\indexisarcmd{def}\indexisarcmd{let}
|
wenzelm@7134
|
423 |
\begin{matharray}{rcl}
|
wenzelm@7134
|
424 |
\isarcmd{fix} & : & \isartrans{proof(state)}{proof(state)} \\
|
wenzelm@7134
|
425 |
\isarcmd{assume} & : & \isartrans{proof(state)}{proof(state)} \\
|
wenzelm@7134
|
426 |
\isarcmd{presume} & : & \isartrans{proof(state)}{proof(state)} \\
|
wenzelm@7134
|
427 |
\isarcmd{def} & : & \isartrans{proof(state)}{proof(state)} \\
|
wenzelm@7134
|
428 |
\isarcmd{let} & : & \isartrans{proof(state)}{proof(state)} \\
|
wenzelm@7134
|
429 |
\end{matharray}
|
wenzelm@7134
|
430 |
|
wenzelm@7134
|
431 |
\begin{rail}
|
wenzelm@7134
|
432 |
'fix' (var +) comment?
|
wenzelm@7134
|
433 |
;
|
wenzelm@7134
|
434 |
('assume' | 'presume') thmdecl? (proppat +) comment?
|
wenzelm@7134
|
435 |
;
|
wenzelm@7134
|
436 |
'def' thmdecl? var '==' termpat comment?
|
wenzelm@7134
|
437 |
;
|
wenzelm@7134
|
438 |
'let' ((term + 'as') '=' term comment? + 'and')
|
wenzelm@7134
|
439 |
;
|
wenzelm@7134
|
440 |
|
wenzelm@7134
|
441 |
var: name ('::' type)?
|
wenzelm@7134
|
442 |
;
|
wenzelm@7134
|
443 |
\end{rail}
|
wenzelm@7134
|
444 |
|
wenzelm@7167
|
445 |
\begin{descr}
|
wenzelm@7167
|
446 |
\item [$\FIX{x}$] FIXME
|
wenzelm@7167
|
447 |
\item [$\ASSUME{a}{\Phi}$ and $\PRESUME{a}{\Phi}$] FIXME
|
wenzelm@7167
|
448 |
\item [$\DEF{a}{x \equiv t}$] FIXME
|
wenzelm@7167
|
449 |
\item [$\LET{\vec p = \vec t}$] FIXME
|
wenzelm@7167
|
450 |
\end{descr}
|
wenzelm@7167
|
451 |
|
wenzelm@7167
|
452 |
|
wenzelm@7167
|
453 |
\subsection{Facts and forward chaining}
|
wenzelm@7167
|
454 |
|
wenzelm@7167
|
455 |
FIXME
|
wenzelm@7167
|
456 |
|
wenzelm@7167
|
457 |
\indexisarcmd{note}\indexisarcmd{then}\indexisarcmd{from}\indexisarcmd{with}
|
wenzelm@7167
|
458 |
\begin{matharray}{rcl}
|
wenzelm@7167
|
459 |
\isarcmd{note} & : & \isartrans{proof(state)}{proof(state)} \\
|
wenzelm@7167
|
460 |
\isarcmd{then} & : & \isartrans{proof(state)}{proof(chain)} \\
|
wenzelm@7167
|
461 |
\isarcmd{from} & : & \isartrans{proof(state)}{proof(chain)} \\
|
wenzelm@7167
|
462 |
\isarcmd{with} & : & \isartrans{proof(state)}{proof(chain)} \\
|
wenzelm@7167
|
463 |
\end{matharray}
|
wenzelm@7167
|
464 |
|
wenzelm@7167
|
465 |
\begin{rail}
|
wenzelm@7167
|
466 |
'note' thmdef? thmrefs comment?
|
wenzelm@7167
|
467 |
;
|
wenzelm@7167
|
468 |
'then' comment?
|
wenzelm@7167
|
469 |
;
|
wenzelm@7167
|
470 |
('from' | 'with') thmrefs comment?
|
wenzelm@7167
|
471 |
;
|
wenzelm@7167
|
472 |
\end{rail}
|
wenzelm@7167
|
473 |
|
wenzelm@7167
|
474 |
\begin{descr}
|
wenzelm@7167
|
475 |
\item [$\NOTE{a}{bs}$] recalls existing facts $bs$, binding the result as $a$
|
wenzelm@7167
|
476 |
(and $facts$). Note that attributes may be involved as well, both on the
|
wenzelm@7167
|
477 |
left and right hand side.
|
wenzelm@7167
|
478 |
\item [$\THEN$] indicates forward chaining by the current facts in order to
|
wenzelm@7167
|
479 |
establish the subsequent goal. The initial proof method invoked to solve
|
wenzelm@7167
|
480 |
that will be offered these facts to do anything ``appropriate'' (see also
|
wenzelm@7167
|
481 |
\S\ref{sec:proof-steps}). For example, method $rule$ (see
|
wenzelm@7167
|
482 |
\S\ref{sec:pure-meth}) would do an elimination rather than an introduction.
|
wenzelm@7167
|
483 |
\item [$\FROM{bs}$] abbreviates $\NOTE{facts}{bs}~\THEN$; also note that
|
wenzelm@7167
|
484 |
$\THEN$ is equivalent to $\FROM{facts}$.
|
wenzelm@7167
|
485 |
\item [$\WITH{bs}$] abbreviates $\FROM{bs~facts}$; thus the forward chaining
|
wenzelm@7167
|
486 |
is from earlier facts together with the current ones.
|
wenzelm@7167
|
487 |
\end{descr}
|
wenzelm@7167
|
488 |
|
wenzelm@7167
|
489 |
|
wenzelm@7167
|
490 |
\subsection{Goal statements}
|
wenzelm@7167
|
491 |
|
wenzelm@7167
|
492 |
Proof mode is entered from theory mode by initial goal commands $\THEOREMNAME$
|
wenzelm@7167
|
493 |
and $\LEMMANAME$. New local goals may be claimed within proof mode: four
|
wenzelm@7167
|
494 |
variants indicate whether the result is meant to solve some pending goal and
|
wenzelm@7167
|
495 |
whether forward chaining is employed.
|
wenzelm@7167
|
496 |
|
wenzelm@7167
|
497 |
\indexisarcmd{theorem}\indexisarcmd{lemma}
|
wenzelm@7167
|
498 |
\indexisarcmd{have}\indexisarcmd{show}\indexisarcmd{hence}\indexisarcmd{thus}
|
wenzelm@7167
|
499 |
\begin{matharray}{rcl}
|
wenzelm@7167
|
500 |
\isarcmd{theorem} & : & \isartrans{theory}{proof(prove)} \\
|
wenzelm@7167
|
501 |
\isarcmd{lemma} & : & \isartrans{theory}{proof(prove)} \\
|
wenzelm@7167
|
502 |
\isarcmd{have} & : & \isartrans{proof(state)}{proof(prove)} \\
|
wenzelm@7167
|
503 |
\isarcmd{show} & : & \isartrans{proof(state)}{proof(prove)} \\
|
wenzelm@7167
|
504 |
\isarcmd{hence} & : & \isartrans{proof(state)}{proof(prove)} \\
|
wenzelm@7167
|
505 |
\isarcmd{thus} & : & \isartrans{proof(state)}{proof(prove)} \\
|
wenzelm@7167
|
506 |
\end{matharray}
|
wenzelm@7167
|
507 |
|
wenzelm@7167
|
508 |
\begin{rail}
|
wenzelm@7167
|
509 |
('theorem' | 'lemma') goal
|
wenzelm@7167
|
510 |
;
|
wenzelm@7167
|
511 |
('have' | 'show' | 'hence' | 'thus') goal
|
wenzelm@7167
|
512 |
;
|
wenzelm@7167
|
513 |
|
wenzelm@7167
|
514 |
goal: thmdecl? proppat comment?
|
wenzelm@7167
|
515 |
;
|
wenzelm@7167
|
516 |
\end{rail}
|
wenzelm@7167
|
517 |
|
wenzelm@7167
|
518 |
\begin{descr}
|
wenzelm@7167
|
519 |
\item [$\THEOREM{name}{\phi}$] enters proof mode with $\phi$ as main goal,
|
wenzelm@7167
|
520 |
eventually resulting in some theorem $\turn \phi$, which stored in the
|
wenzelm@7167
|
521 |
theory.
|
wenzelm@7167
|
522 |
\item [$\LEMMANAME$] is similar to $\THEOREMNAME$, but tags the result as
|
wenzelm@7167
|
523 |
``lemma''.
|
wenzelm@7167
|
524 |
\item [$\HAVE{name}{\phi}$] claims a local goal, eventually resulting in a
|
wenzelm@7167
|
525 |
theorem with the current assumption context as hypotheses.
|
wenzelm@7167
|
526 |
\item [$\SHOW{name}{\phi}$] same as $\HAVE{name}{\phi}$, but solves some
|
wenzelm@7167
|
527 |
pending goal with the result exported to the enclosing assumption context.
|
wenzelm@7167
|
528 |
\item [$\HENCE{name}{\phi}$] abbreviates $\THEN~\HAVE{name}{\phi}$, i.e.\
|
wenzelm@7167
|
529 |
claims a local goal to be proven by forward chaining the current facts.
|
wenzelm@7167
|
530 |
\item [$\THUS{name}{\phi}$] abbreviates $\THEN~\SHOW{name}{\phi}$.
|
wenzelm@7167
|
531 |
\end{descr}
|
wenzelm@7167
|
532 |
|
wenzelm@7167
|
533 |
|
wenzelm@7167
|
534 |
\subsection{Initial and terminal proof steps}\label{sec:proof-steps}
|
wenzelm@7167
|
535 |
|
wenzelm@7167
|
536 |
Arbitrary goal refinements via tactics is considered harmful. Consequently
|
wenzelm@7167
|
537 |
the Isar framework admits proof methods to be invoked in two places only.
|
wenzelm@7167
|
538 |
\begin{enumerate}
|
wenzelm@7167
|
539 |
\item An \emph{initial} refinement step (via $\PROOF{m@1}$) reduces a newly
|
wenzelm@7167
|
540 |
stated intermediate goal to a number of sub-goals that are to be solved
|
wenzelm@7167
|
541 |
subsequently. Facts are passed to $m@1$ for forward chaining if so
|
wenzelm@7167
|
542 |
indicated by $proof(chain)$ mode.
|
wenzelm@7167
|
543 |
|
wenzelm@7167
|
544 |
\item A \emph{terminal} conclusion step (via $\QED{m@2}$)) solves any remaining
|
wenzelm@7167
|
545 |
pending goals completely. No facts are passed to $m@2$.
|
wenzelm@7167
|
546 |
\end{enumerate}
|
wenzelm@7167
|
547 |
|
wenzelm@7167
|
548 |
The only other proper way to affect pending goals is by $\SHOWNAME$, which
|
wenzelm@7167
|
549 |
involves an explicit statement of what is solved.
|
wenzelm@7167
|
550 |
|
wenzelm@7167
|
551 |
Also note that initial proof methods should either solve the goal completely,
|
wenzelm@7167
|
552 |
or constitute some well-understood deterministic reduction to new sub-goals.
|
wenzelm@7167
|
553 |
Arbitrary automatic proof tools that are prone leave a large number of badly
|
wenzelm@7167
|
554 |
structured sub-goals are no help in continuing the proof document in any
|
wenzelm@7167
|
555 |
intelligible way. A much better technique is to $\SHOWNAME$ some non-trivial
|
wenzelm@7167
|
556 |
reduction as an explicit rule, which is solved completely by some automated
|
wenzelm@7167
|
557 |
method, and then applied to some pending goal.
|
wenzelm@7167
|
558 |
|
wenzelm@7167
|
559 |
\indexisarcmd{proof}\indexisarcmd{qed}\indexisarcmd{by}
|
wenzelm@7167
|
560 |
\indexisarcmd{.}\indexisarcmd{..}\indexisarcmd{sorry}
|
wenzelm@7167
|
561 |
\begin{matharray}{rcl}
|
wenzelm@7167
|
562 |
\isarcmd{proof} & : & \isartrans{proof(prove)}{proof(state)} \\
|
wenzelm@7167
|
563 |
\isarcmd{qed} & : & \isartrans{proof(state)}{proof(state) ~|~ theory} \\
|
wenzelm@7167
|
564 |
\isarcmd{by} & : & \isartrans{proof(prove)}{proof(state) ~|~ theory} \\
|
wenzelm@7167
|
565 |
\isarcmd{..} & : & \isartrans{proof(prove)}{proof(state) ~|~ theory} \\
|
wenzelm@7167
|
566 |
\isarcmd{.} & : & \isartrans{proof(prove)}{proof(state) ~|~ theory} \\
|
wenzelm@7167
|
567 |
\isarcmd{sorry} & : & \isartrans{proof(prove)}{proof(state) ~|~ theory} \\
|
wenzelm@7167
|
568 |
\end{matharray}
|
wenzelm@7167
|
569 |
|
wenzelm@7167
|
570 |
\begin{rail}
|
wenzelm@7167
|
571 |
'proof' interest? meth? comment?
|
wenzelm@7167
|
572 |
;
|
wenzelm@7167
|
573 |
'qed' meth? comment?
|
wenzelm@7167
|
574 |
;
|
wenzelm@7167
|
575 |
'by' meth meth? comment?
|
wenzelm@7167
|
576 |
;
|
wenzelm@7167
|
577 |
('.' | '..' | 'sorry') comment?
|
wenzelm@7167
|
578 |
;
|
wenzelm@7167
|
579 |
|
wenzelm@7167
|
580 |
meth: method interest?
|
wenzelm@7167
|
581 |
;
|
wenzelm@7167
|
582 |
\end{rail}
|
wenzelm@7167
|
583 |
|
wenzelm@7167
|
584 |
\begin{descr}
|
wenzelm@7167
|
585 |
\item [$\PROOF{m}$] refines the pending goal by proof method $m$ (facts for
|
wenzelm@7167
|
586 |
forward chaining are passed if indicated by $proof(chain)$).
|
wenzelm@7167
|
587 |
\item [$\QED{m}$] refines any remaining goals by proof method $m$ and
|
wenzelm@7167
|
588 |
concludes the sub-proof. If the goal had been $\SHOWNAME$, some pending
|
wenzelm@7167
|
589 |
sub-goal is solved as well by the rule resulting from the result exported to
|
wenzelm@7167
|
590 |
the enclosing goal context.
|
wenzelm@7167
|
591 |
|
wenzelm@7167
|
592 |
Thus $\QEDNAME$ may fail for two reasons: either $m$ fails to solve all
|
wenzelm@7167
|
593 |
remaining goals completely, or the resulting rule does not resolve with any
|
wenzelm@7167
|
594 |
enclosing goal. Debugging such a situation might involve temporarily
|
wenzelm@7167
|
595 |
changing $\SHOWNAME$ into $\HAVENAME$, or weakening the local context by
|
wenzelm@7167
|
596 |
replacing $\ASSUMENAME$ by $\PRESUMENAME$.
|
wenzelm@7167
|
597 |
\item [$\BY{m@1}{m@2}$] is a \emph{terminal proof}; it abbreviates
|
wenzelm@7167
|
598 |
$\PROOF{m@1}~\QED{m@2}$, automatically backtracking across both methods.
|
wenzelm@7167
|
599 |
|
wenzelm@7167
|
600 |
Debugging an unsuccessful $\BY{m@1}{m@2}$ commands might be done by simply
|
wenzelm@7167
|
601 |
expanding the abbreviation by hand; usually $\PROOF{m@1}$ is already
|
wenzelm@7167
|
602 |
sufficient to see what goes wrong.
|
wenzelm@7167
|
603 |
\item [$\isarkeyword{..}$] is a \emph{default proof}; it abbreviates
|
wenzelm@7167
|
604 |
$\BY{default}{finish}$, where method $default$ usually applies a single
|
wenzelm@7167
|
605 |
elimination or introduction rule according to the topmost symbol, and
|
wenzelm@7167
|
606 |
$finish$ solves all goals by assumption.
|
wenzelm@7167
|
607 |
\item [$\isarkeyword{.}$] is a \emph{trivial proof}, it abbreviates
|
wenzelm@7167
|
608 |
$\BY{-}{finish}$, where method $-$ does nothing except inserting any facts
|
wenzelm@7167
|
609 |
into the proof state.
|
wenzelm@7167
|
610 |
\item [$\isarkeyword{sorry}$] is a \emph{fake proof}; provided that
|
wenzelm@7167
|
611 |
\texttt{quick_and_dirty} is enabled, $\isarkeyword{sorry}$ pretends to solve
|
wenzelm@7167
|
612 |
the goal without much ado. Of course, the result is a fake theorem only,
|
wenzelm@7167
|
613 |
involving some oracle in its internal derivation object. Note that this is
|
wenzelm@7167
|
614 |
indicated as \texttt{[!]} in the printed result. The main application of
|
wenzelm@7167
|
615 |
$\isarkeyword{sorry}$ is to support top-down proof development.
|
wenzelm@7167
|
616 |
\end{descr}
|
wenzelm@7134
|
617 |
|
wenzelm@7134
|
618 |
|
wenzelm@7134
|
619 |
\subsection{Block structure}
|
wenzelm@7134
|
620 |
|
wenzelm@7167
|
621 |
While Isar is inherently block-structured, opening and closing blocks is
|
wenzelm@7167
|
622 |
mostly handled rather casually, with little explicit user-intervention. Any
|
wenzelm@7167
|
623 |
local goal statement automatically opens \emph{two} blocks, which are closed
|
wenzelm@7167
|
624 |
again when concluding the sub-proof (by $\QEDNAME$ etc.). Sections of
|
wenzelm@7167
|
625 |
different context within a sub-proof are typically switched via
|
wenzelm@7167
|
626 |
$\isarkeyword{next}$, which is just a single block-close followed by
|
wenzelm@7167
|
627 |
block-open again. Thus the effect of $\isarkeyword{next}$ is to reset the
|
wenzelm@7167
|
628 |
proof context to that of the head of the sub-proof. Note that there is no
|
wenzelm@7167
|
629 |
goal focus involved!
|
wenzelm@7167
|
630 |
|
wenzelm@7167
|
631 |
There are explicit block parentheses as well. These typically achieve a
|
wenzelm@7167
|
632 |
strong forward style of reasoning.
|
wenzelm@7167
|
633 |
|
wenzelm@7134
|
634 |
\indexisarcmd{next}\indexisarcmd{\{\{}\indexisarcmd{\}\}}
|
wenzelm@7134
|
635 |
\begin{matharray}{rcl}
|
wenzelm@7134
|
636 |
\isarcmd{next} & : & \isartrans{proof(state)}{proof(state)} \\
|
wenzelm@7134
|
637 |
\isarcmd{\{\{} & : & \isartrans{proof(state)}{proof(state)} \\
|
wenzelm@7134
|
638 |
\isarcmd{\}\}} & : & \isartrans{proof(state)}{proof(state)} \\
|
wenzelm@7134
|
639 |
\end{matharray}
|
wenzelm@7134
|
640 |
|
wenzelm@7167
|
641 |
\begin{descr}
|
wenzelm@7167
|
642 |
\item [$\isarkeyword{next}$] switches to a fresh block within a sub-proof,
|
wenzelm@7167
|
643 |
resetting the context to the initial one.
|
wenzelm@7167
|
644 |
\item [$\isarkeyword{\{\{}$ and $\isarkeyword{\}\}}$] explicitly open and
|
wenzelm@7167
|
645 |
close blocks. Any current facts pass through $\isarkeyword{\{\{}$
|
wenzelm@7167
|
646 |
unchanged, while $\isarkeyword{\}\}}$ causes them to be \emph{exported} into
|
wenzelm@7167
|
647 |
the enclosing context. Thus fixed variables are generalized, assumptions
|
wenzelm@7167
|
648 |
discharged, and local definitions eliminated.
|
wenzelm@7167
|
649 |
\end{descr}
|
wenzelm@7134
|
650 |
|
wenzelm@7134
|
651 |
\subsection{Calculational proof}
|
wenzelm@7134
|
652 |
|
wenzelm@7134
|
653 |
\indexisarcmd{also}\indexisarcmd{finally}
|
wenzelm@7134
|
654 |
\begin{matharray}{rcl}
|
wenzelm@7134
|
655 |
\isarcmd{also} & : & \isartrans{proof(state)}{proof(state)} \\
|
wenzelm@7134
|
656 |
\isarcmd{finally} & : & \isartrans{proof(state)}{proof(chain)} \\
|
wenzelm@7134
|
657 |
\end{matharray}
|
wenzelm@7134
|
658 |
|
wenzelm@7134
|
659 |
\begin{rail}
|
wenzelm@7134
|
660 |
('also' | 'finally') transrules? comment?
|
wenzelm@7134
|
661 |
;
|
wenzelm@7134
|
662 |
|
wenzelm@7134
|
663 |
transrules: '(' thmrefs ')' interest?
|
wenzelm@7134
|
664 |
;
|
wenzelm@7134
|
665 |
\end{rail}
|
wenzelm@7134
|
666 |
|
wenzelm@7167
|
667 |
\begin{descr}
|
wenzelm@7134
|
668 |
\item [$ $] FIXME
|
wenzelm@7167
|
669 |
\end{descr}
|
wenzelm@7134
|
670 |
|
wenzelm@7134
|
671 |
|
wenzelm@7134
|
672 |
|
wenzelm@7134
|
673 |
\subsection{Improper proof steps}
|
wenzelm@7134
|
674 |
|
wenzelm@7167
|
675 |
The following commands emulate tactic scripts to some extent. While these are
|
wenzelm@7167
|
676 |
anathema for writing proper Isar proof documents, they might come in handy for
|
wenzelm@7167
|
677 |
exploring and debugging.
|
wenzelm@7167
|
678 |
|
wenzelm@7167
|
679 |
\indexisarcmd{apply}\indexisarcmd{then-apply}\indexisarcmd{back}
|
wenzelm@7134
|
680 |
\begin{matharray}{rcl}
|
wenzelm@7134
|
681 |
\isarcmd{apply}^* & : & \isartrans{proof}{proof} \\
|
wenzelm@7134
|
682 |
\isarcmd{then_apply}^* & : & \isartrans{proof}{proof} \\
|
wenzelm@7134
|
683 |
\isarcmd{back}^* & : & \isartrans{proof}{proof} \\
|
wenzelm@7134
|
684 |
\end{matharray}
|
wenzelm@7134
|
685 |
|
wenzelm@7134
|
686 |
\railalias{thenapply}{then\_apply}
|
wenzelm@7134
|
687 |
\railterm{thenapply}
|
wenzelm@7134
|
688 |
|
wenzelm@7134
|
689 |
\begin{rail}
|
wenzelm@7134
|
690 |
'apply' method
|
wenzelm@7134
|
691 |
;
|
wenzelm@7134
|
692 |
thenapply method
|
wenzelm@7134
|
693 |
;
|
wenzelm@7134
|
694 |
'back'
|
wenzelm@7134
|
695 |
;
|
wenzelm@7134
|
696 |
\end{rail}
|
wenzelm@7134
|
697 |
|
wenzelm@7167
|
698 |
\begin{descr}
|
wenzelm@7134
|
699 |
\item [$ $] FIXME
|
wenzelm@7167
|
700 |
\end{descr}
|
wenzelm@7134
|
701 |
|
wenzelm@7134
|
702 |
|
wenzelm@7134
|
703 |
\section{Other commands}
|
wenzelm@7134
|
704 |
|
wenzelm@7134
|
705 |
\subsection{Diagnostics}
|
wenzelm@7134
|
706 |
|
wenzelm@7134
|
707 |
\indexisarcmd{typ}\indexisarcmd{term}\indexisarcmd{prop}\indexisarcmd{thm}
|
wenzelm@7134
|
708 |
\begin{matharray}{rcl}
|
wenzelm@7134
|
709 |
\isarcmd{typ} & : & \isarkeep{theory~|~proof} \\
|
wenzelm@7134
|
710 |
\isarcmd{term} & : & \isarkeep{theory~|~proof} \\
|
wenzelm@7134
|
711 |
\isarcmd{prop} & : & \isarkeep{theory~|~proof} \\
|
wenzelm@7134
|
712 |
\isarcmd{thm} & : & \isarkeep{theory~|~proof} \\
|
wenzelm@7134
|
713 |
\end{matharray}
|
wenzelm@7134
|
714 |
|
wenzelm@7134
|
715 |
\begin{rail}
|
wenzelm@7134
|
716 |
'typ' type
|
wenzelm@7134
|
717 |
;
|
wenzelm@7134
|
718 |
'term' term
|
wenzelm@7134
|
719 |
;
|
wenzelm@7134
|
720 |
'prop' prop
|
wenzelm@7134
|
721 |
;
|
wenzelm@7134
|
722 |
'thm' thmrefs
|
wenzelm@7134
|
723 |
;
|
wenzelm@7134
|
724 |
\end{rail}
|
wenzelm@7134
|
725 |
|
wenzelm@7167
|
726 |
\begin{descr}
|
wenzelm@7134
|
727 |
\item [$\isarkeyword{typ}~\tau$, $\isarkeyword{term}~t$,
|
wenzelm@7134
|
728 |
$\isarkeyword{prop}~\phi$] read and print types / terms / propositions
|
wenzelm@7134
|
729 |
according to the current theory or proof context.
|
wenzelm@7134
|
730 |
\item [$\isarkeyword{thm}~thms$] retrieves lists of theorems from the current
|
wenzelm@7134
|
731 |
theory or proof context. Note that any attributes included in the theorem
|
wenzelm@7134
|
732 |
specifications are applied to a temporary proof context derived from the
|
wenzelm@7134
|
733 |
current theory or proof; the resulting context is discarded.
|
wenzelm@7167
|
734 |
\end{descr}
|
wenzelm@7134
|
735 |
|
wenzelm@7134
|
736 |
|
wenzelm@7134
|
737 |
\subsection{System operations}
|
wenzelm@7134
|
738 |
|
wenzelm@7167
|
739 |
\indexisarcmd{cd}\indexisarcmd{pwd}\indexisarcmd{use-thy}\indexisarcmd{use-thy-only}
|
wenzelm@7167
|
740 |
\indexisarcmd{update-thy}\indexisarcmd{update-thy-only}
|
wenzelm@7134
|
741 |
\begin{matharray}{rcl}
|
wenzelm@7134
|
742 |
\isarcmd{cd} & : & \isarkeep{\cdot} \\
|
wenzelm@7134
|
743 |
\isarcmd{pwd} & : & \isarkeep{\cdot} \\
|
wenzelm@7134
|
744 |
\isarcmd{use_thy} & : & \isarkeep{\cdot} \\
|
wenzelm@7134
|
745 |
\isarcmd{use_thy_only} & : & \isarkeep{\cdot} \\
|
wenzelm@7134
|
746 |
\isarcmd{update_thy} & : & \isarkeep{\cdot} \\
|
wenzelm@7134
|
747 |
\isarcmd{update_thy_only} & : & \isarkeep{\cdot} \\
|
wenzelm@7134
|
748 |
\end{matharray}
|
wenzelm@7134
|
749 |
|
wenzelm@7167
|
750 |
\begin{descr}
|
wenzelm@7134
|
751 |
\item [$\isarkeyword{cd}~name$] changes the current directory of the Isabelle
|
wenzelm@7134
|
752 |
process.
|
wenzelm@7134
|
753 |
\item [$\isarkeyword{pwd}~$] prints the current working directory.
|
wenzelm@7134
|
754 |
\item [$\isarkeyword{use_thy}~name$, $\isarkeyword{use_thy_only}~name$,
|
wenzelm@7134
|
755 |
$\isarkeyword{update_thy}~name$, $\isarkeyword{update_thy_only}~name$] load
|
wenzelm@7134
|
756 |
theory files. These commands are exactly the same as the corresponding ML
|
wenzelm@7134
|
757 |
functions (see also \cite[\S1 and \S6]{isabelle-ref}).
|
wenzelm@7167
|
758 |
\end{descr}
|
wenzelm@7134
|
759 |
|
wenzelm@7134
|
760 |
|
wenzelm@7046
|
761 |
%%% Local Variables:
|
wenzelm@7046
|
762 |
%%% mode: latex
|
wenzelm@7046
|
763 |
%%% TeX-master: "isar-ref"
|
wenzelm@7046
|
764 |
%%% End:
|