2 Installation and Download locations

This section explains the installation for different operating systems (OSs). As iTS MAITis heavily based on the open source software packages Python2 (http://www.python.org), Python Numeric and Tcl/Tk (http://dev.scriptics.com/), it is necessary to install those packages before iTS MAITcan be launched¹.

DIRECT LINKS FOR DOWNLOADING ALL REQUIRED SOFTWARE ARE PROVIDED IN THE SUBSECTION THAT CORRESPOND TO YOUR OPERATING SYSTEM!. For windows, there is just one zipped file to download, which contains all the software needed!

2.1 System requirements

As system requirement, an Intel Processor with at least 500MHz (or equivalent), 64Mb of RAM and 15MB of disk space is recommended. The simulator can also be used with less powerful computers. However, in the graphic mode, vehicles movement becomes very slow and jerky. However, a previously edited network can be simulated in command-line mode, which is significantly faster.

2.2 Installation for MS Windows

All free packages run on 95, 98, NT, 2000, XP and Vista.

2.2.1 Instructions if open source packages are not installed

Since code.enthought.com offers a complete Python suit for windows, including all required packages, it is easiest to install just this software (instead of the individual packages):

1. Download the latest version at http://code.enthought.com/enthon/.

2. Double-click on the downloaded file and follow instructions.

Now you are ready to install the iTS MAITsoftware, (see Sec. 2.2.2).

2.2.2 Open source packages are already installed

Use these instructions if you have Python 2.3.4 and Python Numeric-23.1 or greater installed on your system (or followed installation in Sec. 2.2.1).

1. Download
   http://distart041.ing.unibo.it/~mait/projects/sim/downloads/its-0.51-win.zip into a directory of your choice i.e. My Documents its. On this level, a directory (containing all simulation files) with the name its-0.51will be created.

2. Rightclick on file My Documents its and select: extract all...

3. Open the its-0.51directory and start the simulator by clicking on the file its (with the snake icon). If you can see the main window of the simulator (see Fig. 2), you are ready to proceed with the tutorials in Sec. 3.

Remark:

(ii): If you have already used a previous version of iTS MAITyou may have also previous version of Python and Numeric. These versions should work as well. However, Python 2.3.4 has been a major bug-fix release and runs indeed much more stable and also faster. So if you intend to make large simulations with several thousands of users you are encouraged to reinstall all packages. Then follow the instructions of Sec. 2.2.1.

(ii): Several iTS MAIT-version can coexist in separate directory trees, always named by its version. Executing the simulator of one or the other version is done by diving into the respective directory and clicking on the file its. The simulation files created be an older version (usually in the ``projects'' sub-folder), can be copied into the ``projects'' sub-folder of the newer version. Simulation files created by an older version, will be automatically upgraded when opened with the newer version of the simulator.

2.2.3 Installation of PyOpenGL

In order to view your network in 3D and real-time, you need to have the PyOpenGl-package installed (!! Does not work with Vista - sorry):

1. install one more Python module: download exactly this file PyOpenGL-2.0.1.09.py2.3-numpy23.exe (for Python 2.3) or PyOpenGL-2.0.1.09.py2.4-numpy23.exe (for Python 2.4). from
   
   
   and install this self-extracting file by double-clicking and following instructions.

2. Then you need the Microsoft's OpenGL driver, which is no Open Source but you get the binary for free from Microsoft or an open-source clone at
   http://distart041.ing.unibo.it/~mait/projects/sim/downloads/glut-3.7.6-bin.zip
   Unzip the file, and browse into the directory. You will see a file called glut32.dll. Just copy this dll- file in the main directory of the previously installed OpenGL Python package which is usually in
   
   .

2.3 Installation for Linux

Linux installation is in the best case very easy because Python is part of most Linux distributions. If you have all the packages or if you have installed a previous version of iTS MAITyou can directly go to Sec. 2.3.1.

However, for the same reason it can get messy if you have some packages installed, other not, or if you are using older packages. Therefore, first check out whether the above mentioned open source packages are already installed on your system:

1. Open a terminal and type:python, or python2. If you get an error message or if you have a version lower than 2.2 then you need to install/upgrade python and all other packages. Go to Sec. 2.3.2.

2. If you have a valid version of python then type: import Tkinter behind the python prompt. If this produces an error message, you need to install TKinter.Go to Sec. 2.3.2.

3. If you have the right version of python then type import Numeric after the python prompt. If this produces an error message, you need to install Python Numeric. Go to Sec. 2.3.3.

Some quick installation method for python and the required packages are briefly described below. If you have all packages you can continue with Sec. 2.3.1.

2.3.1 Installing iTS MAIT

Once all packages are in place the installation of iTS MAITis easy:

1. Download
   http://distart041.ing.unibo.it/~mait/projects/sim/downloads/its-0.51.zip.
2. place file its-0.51.zip in a directory of your choice. Open a Terminal, open this directory and unpack zip archive with $ unzip -a its-0.51.zip.

3. A new subdirectory called ``its-0.51'' has been created; open it and start the simulator with$ python its.py. For command line option type $ python its.py -h. If you see the main window of the simulator (see Fig. 2) then the installation is successfully terminated. Otherwise, check out the messages on the terminal to see whether there is a package missing.

Remark 1: If you intend to install iTS MAITfor all users then you should put the entire its-0.51directory in $PYTHONHOME/site-packages/. Read the documentation on how to make the new package visible to the Python interpreter.

Remark 2: If you intend do more Python programming, I recommend to use the Python text editor (and development tool) IDLE (which you can find in the Python installation tree if it is installed). There are also syntax highlighting packages for the Emacs and FTE editors. All tools can be found at www.python.org.

2.3.2 Installing/upgrading Python and Tkinter

What you need is the version Python 2.2 or higher (2.3 recommended) with Tkinter support. There are different methods to install/upgrade which are more or less easy, dependent on your present system configuration.

• Installing with CDs of distribution: This is easiest. Since Python and Tkinter are part of any Linux distribution that I know of. Take these CDs and install the following packages in this sequence (if not already installed):
  1. python-2.3.3-6
  2. python-devel-2.3.3-6 (only if you have no RPM binary for Numeric)
  3. tcl-8.4.5-7 (required for Tkinter)
  4. tk-8.4.5-8 (required for Tkinter)
  5. tix-8.1.4-96.1 (required for Tkinter)
  6. tkinter-2.3.3-6

  This list is made off the Fedora Core 2 distribution. In other distributions version names may slightly differ, but the important thing is that they are consistent and satisfy internal dependencies.

  RPM package installation for Linux newcomers: login as root, open a terminal, cd into the RPM-directory of the distribution CD (usually: $ cd /mnt/cdrom/RPMS) and type:
  # rpm -i packagename.rpm
  Alternatively use your favorite graphical RPM management application which is usually somewhere in the applications launch-bar under System tools, Admin tools or similar).

• Upgrading: If you have Python version $>=$  2.2 the current iTS MAITversion should work. However, Python 2.3 has been a major bug-fix release and runs indeed much more stable and also faster (up to 30% on some benchmarks). So if you intend to make large simulations with several thousands of users you are encouraged to upgrade Python. Several Python versions can actually coexist on your system, you just need to make sure which one is called when you type $ python or click on a python file.

  If you have an RPM based Linux system, make sure to get the upgrade RPMs for your specific distribution, otherwise you will get crazy with package dependencies. You can find upgrades on one of the following sites:
  http://www.rpmfind.net/linux/RPM/(all major distributions),
  http://freshrpms.net/packages/(RedHat oriented),
  http://dag.wieers.com/home-made/apt/(RedHat oriented).
  Other distributions like Debian or Gentoo do maintain their own database with upgrades, have a look at their home page.

• Compile from source: Compiled code usually runs faster than binaries because it will be compiled for your processor and not for the (lowest) standard processor i386. Download the source code for Python 2.3.4 from the download page at www.python.org/ftp/python/2.3.4/Python-2.3.4.tgz and compile. Even though the compilation process is well done and straight forward to use, people who never compiled a source package may find it difficult to go through all steps.

2.3.3 Installing Python Numeric

If you have Python up and running, choose one of these methods to get python-numeric-23.1 installed:

• RPMs: This is easiest, but recommended ONLY if you find the package python-numeric-23.1-xxxx.rpm for your particular Linux distribution. Have a look at the following mirrors: www.rpmfind.net/linux/RPM/(all major distributions), http://freshrpms.net/packages/(RedHat oriented), http://dag.wieers.com/home-made/apt/(RedHat oriented) Other distributions like Debian or Gentoo do maintain their own database with upgrades. Just have a look at the respective home page.

• Compile from source: This is actually the ``Python way'' to install Python modules, because it is platform independent.
  1. First make sure you have the Python header files installed, which should be the case when you have compiled Python from source. Otherwise,install the RPM package python-devel-2.3.x-x from your Linux distribution.

  2. Download source Numeric-23.1.zip at
     http://sourceforge.net/projects/numpy and unzip.

  3. Open terminal as root in the unzipped directory Numeric-23.1 and type:
     # python setup.py install
     
  If you have problems consult the Python Numeric homepage (http://www.pfdubois.com/numpy/)

2.3.4 Installation of PyOpenGL

In order to view your network in 3D and real-time, you need to have the PyOpenGL-2.0 (or newer) installed:

• PyOpenGL is often available as binary RPM or Debian package for you Linux distribution. Sometime the package is called ``python-opengl''. Check out the respective repositories, see Sec. 2.3.3

• Download and compile the source from