~~ This describes how to run HOP inside a Java Virtual Machine. Although
it is still possible to compile HOP to JVM bytecodes, we no longer
distribute pre-compiled JVM version of HOP. This manual is then
obsolete and most applies to old HOP versions.

== 1. Introduction ==

~~ HOP can either be executed in native mode or in JVM mode. The native
mode is only operational on Unix systems. The pre-compiled JVM mode
is available on all computers equipped with a complete Java
Development Kit 1.4 or greater. In particular, this port is operational
on Linux, MacOS X, and Microsoft Windows.

~~ As much as possible the native version should be prefered to the JVM
pre-compiled distribution. The JVM pre-compiled distribution is still
experimental.

~~ The pre-compiled HOP version is slower that the native version (about 
a factor of 4) and the database library (++sqlite++) is only partially 
implemented.

== 2. Requirements ==

An operational JDK 1.4.2 or higher is needed.

== 3. Executing ==

To start HOP, simply launch the jvm with the ++hop.jar++ file. This
operation is machine dependent. On Unix system, the command looks like:

    java -jar hop-<VERSION>.jar

== 4. Compiling HOP for the JVM ==

~~ For compiling HOP to JVM bytecodes, follows the procedure described in 
[[/hop/home/install|HOP Installation page]] with one exception: one 
configure option must be added. For the JVM backend, HOP must be configured
as:

    ./configure --backend=jvm

 or
    ./configure --backend=jvm --prefix=<your-prefix>