author | urbanc |
Tue, 08 Feb 2011 19:54:23 +0000 | |
changeset 83 | f438f4dbaada |
parent 82 | 14b12b5de6d3 |
child 88 | 1436fc451bb9 |
permissions | -rw-r--r-- |
24 | 1 |
\documentclass{llncs} |
2 |
\usepackage{isabelle} |
|
3 |
\usepackage{isabellesym} |
|
4 |
\usepackage{amsmath} |
|
5 |
\usepackage{amssymb} |
|
6 |
\usepackage{tikz} |
|
7 |
\usepackage{pgf} |
|
8 |
\usepackage{pdfsetup} |
|
9 |
\usepackage{ot1patch} |
|
10 |
\usepackage{times} |
|
11 |
\usepackage{proof} |
|
52
4a517c6ac07d
tuning of the syntax; needs the stmaryrd latex package
urbanc
parents:
24
diff
changeset
|
12 |
\usepackage{stmaryrd} |
24 | 13 |
|
14 |
\urlstyle{rm} |
|
15 |
\isabellestyle{it} |
|
16 |
\renewcommand{\isastyleminor}{\it}% |
|
17 |
\renewcommand{\isastyle}{\normalsize\it}% |
|
18 |
||
19 |
||
20 |
\def\dn{\,\stackrel{\mbox{\scriptsize def}}{=}\,} |
|
21 |
\renewcommand{\isasymequiv}{$\dn$} |
|
22 |
\renewcommand{\isasymemptyset}{$\varnothing$} |
|
23 |
\renewcommand{\isacharunderscore}{\mbox{$\_\!\_$}} |
|
24 |
||
83 | 25 |
\newcommand{\isasymcalL}{\ensuremath{\cal{L}}} |
24 | 26 |
\begin{document} |
27 |
||
54 | 28 |
\title{A Formalisation of the Myhill-Nerode Theorem\\ based on Regular |
29 |
Expressions (Proof Pearl)} |
|
24 | 30 |
\author{Chunhan Wu\inst{1} \and Xingjuan Zhang\inst{1} \and Christian Urban\inst{2}} |
31 |
\institute{PLA University, China \and TU Munich, Germany} |
|
32 |
\maketitle |
|
33 |
||
34 |
\begin{abstract} |
|
35 |
There are numerous textbooks on regular languages. Nearly all of them |
|
36 |
introduce the subject by describing finite automata and |
|
37 |
only mentioning on the side a connection with regular expressions. |
|
38 |
Unfortunately, automata are a hassle for formalisations in HOL-based |
|
82 | 39 |
theorem provers. The reason is that they need to be represented as graphs, |
40 |
matrices or functions, none of which are inductive datatypes. Also |
|
60 | 41 |
operations, such as disjoint unions of graphs, are not easily formalisiable |
61 | 42 |
in HOL. In contrast, regular expressions can be defined conveniently |
24 | 43 |
as datatype and a corresponding reasoning infrastructure comes for |
44 |
free. We show in this paper that a central result from formal |
|
45 |
language theory---the Myhill-Nerode theorem---can be recreated |
|
46 |
using only regular expressions. |
|
47 |
\end{abstract} |
|
48 |
||
75 | 49 |
|
24 | 50 |
\input{session} |
51 |
||
52 |
\bibliographystyle{plain} |
|
53 |
\bibliography{root} |
|
54 |
||
55 |
\end{document} |
|
56 |
||
57 |
%%% Local Variables: |
|
58 |
%%% mode: latex |
|
59 |
%%% TeX-master: t |
|
60 |
%%% End: |