<html>
<!-- $Id$ -->
<head>
<title>The Isabelle System Distribution</title>
</head>
<body>
<h1>The Isabelle System Distribution</h1>
<h2>Version information</h2>
This is <strong>Isabelle-94 revision 8</strong> as of May
1997. Basically, it should be backwards compatible to earlier
Isabelle-94 releases. Older versions (like Isabelle-93) are quite
different, though. See also the <tt>NEWS</tt> file in the distribution
for a more detailed list of new features and changes.
<h2>System requirements</h2>
Isabelle requires a real Unix box with sufficient resources. Fun
starts at about 32MB of RAM (somewhat depending on your ML system),
several tens of MB disk space and with a relatively fast CPU.
<p>
Furthermore, it needs the following software, which is not part of the
distribution:
<ul>
<li> A full Standard ML Compiler (e.g. Poly/ML or SML of New Jersey).
<li> The GNU bash shell (version 1.14.x or 2.x).
<li> Perl 5.x - the Pathologically Eclectic Rubbish Lister.
</ul>
The ML system and GNU bash are absolutely essential. Perl is optional
for core functionality, but still highly recommended.
<p>
The following ML system and platform combinations are known to work
quite well:
<ul>
<li> Poly/ML versions 2.x and 3.1 on Suns (running SunOS 4.1.x or
Solaris 2.x).
<li> SML/NJ 0.93 on Suns and SGIs. There seem to be several
problems with Linux and HP-UX.
<li> SML/NJ 1.07 on Suns, Linux, etc.
<li> SML/NJ 1.09.27 on Suns, Linux, etc.
</ul>
<p>
Poly/ML is a commercial product and costs money, but it is stable and
efficient. It requires relatively little memory (starting at about
16MB) and disk space (about 40MB for all distributed object logics).
See <a href="http://www.ahl.co.uk/poly-ml.html">here</a> for more
information.
<p>
SML/NJ needs lots of store and disk space, but it is free. The last
official release is 0.93.
Recent working versions may be better suited, under normal
circumstances. Beware of the many incompatibilities among SML/NJ
versions! From the 1.09.x family, we now only support 1.09.27, or
later.
<p>
To obtain Standard ML of New Jersey, see the <a
href="http://cm.bell-labs.com/cm/cs/what/smlnj/software.html">here</a>. An
unofficial pre-built binary distribution of SML/NJ 1.09.27 for Linux
is available <a
href="http://www4.informatik.tu-muenchen.de/~wenzelm/resources.html">here</a>.
<h2>Installation</h2>
See file <tt>INSTALL</tt> in the Isabelle sources on how to build the
system. Further background information may be found in the
<em>Isabelle System Manual</em>, distributed as <tt>dvi</tt> with the
sources.
<h2>Other sources of information</h2>
<h3>Mailing list</h3>
The electronic mailing list <tt>isabelle-users@cl.cam.ac.uk</tt>
provides a forum for Isabelle users to discuss problems and exchange
information. To join, send a message to
<tt>isabelle-users-request@cl.cam.ac.uk</tt>.
<h3>Personal mail</h3>
<a href="http://www.cl.cam.ac.uk/users/lcp/">Lawrence C Paulson</a><br>
Computer Laboratory<br>
University of Cambridge<br>
Pembroke Street<br>
Cambridge CB2 3QG<br>
England<br>
<br>
E-mail: lcp@cl.cam.ac.uk<br>
Phone: +44-223-334600<br>
Fax: +44-223-334748<br>
<p>
or
<p>
<a href="http://www4.informatik.tu-muenchen.de:80/~nipkow/">Tobias Nipkow</a><br>
Institut fuer Informatik<br>
T. U. Muenchen<br>
D-80290 Muenchen<br>
Germany<br>
<br>
E-mail: nipkow@informatik.tu-muenchen.de<br>
Phone: +49-89-289-22690<br>
Fax: +49-89-289-28183<br>
<p>
<hr>
Please report any problems you encounter. While we shall try to be
helpful, we can accept no responsibility for the deficiences of
Isabelle and their consequences.
<hr>
</body>
</html>