Aquarium 3.0 Alpha 2
Build Date: 18-02-2002

1] What is this?
2] What do I need to make it work?
3] What does 'Alpha' means?
4] What does this Alpha contain?
5] What does previous this Alpha contained?
6] How to use it?
7] Tips
8] I found a bug, what can I do?

*----------------*
| What is this?  |
*----------------*

This the new Aquarium projet. The aim of this project is to 
demonstrate how complex and powerfull genetic algorithm 
can be.

*---------------------------------*
| What do I need to make it work? |
*---------------------------------*
You need a Java Virtual Machine (JVM, also called JRE).
Take the 1.3.1_02 from http://java.sun.com or the 1.4 final.
Note: 1.4 will speed up your simulation by 2.5 (yes 2.5 if you use the GUI)
Note: On my Windows 1.4 and 1.3.1 doesn't live well together
If you have a jdk1.2.x or 1.1.x this won't work.

*--------------------------*
| What does 'Alpha' means? |
*--------------------------*

'Alpha' means that 
- there are lot of things missing
- there are bugs
- there is a very small documentation (user documentation)
- the Java documentation may contains mistakes
- Sources will not be available (some classes will be obfucasted)
- build will not be very user friendly

*-------------------------------*
| What does this Alpha contain? |
*-------------------------------*
In this second Alpha :

1 ] Documentation
    - XML specification
    - Small tutorial for GUI
    - LifeRoom and FoodMap editor documentation
    - The preference.txt file is a commanted preference file
2] For GUI
    - You can now edit/save your preferences using the GUI
    - Editors work better
    - You have direct acces to the log viewer
    - A new status bar will indicates the remaining time and other information
3] For Command line
    - Unvisible things ;-)
    - Code improves

*------------------------------------------*
| What does previous this Alpha contained? |
*------------------------------------------*

- The first Alpha:
you will have acces to 3 different applications.
[1] - No graphics, in this mode you can launch an experiment, 
      but you will see nothing. Logs will be generated, and 
      it will be your only result.
      Advantages: 
        - You can make huge simulation 
        - You will not consume lot of memory
        - You won't be able to interfer with the experiment
        - Should be very stable
      Disadvantages:
        - It's not very funny
        - And all specification must be done throught files  

[2] - With graphics, in this mode you will be able to view what
      is happening.
      Advantages: 
        - You can view every thing
        - You will be able to edit specification throught windows
        - You will be able to interfer with the experiment
        - Is funny to watch
      Disadvantages:
        - Will take much more memory
        - Lots of thing missing right now

[3] - You can visualy analyse your log file with a build in 
      GraphViewer

  
*----------------*
| How to use it? |
*----------------*

[1] commande line.
  - Recommended configuration: P200Mz and 64Mo
  - Take a look at the preferences.ini file
  - Use the runcmd.sh or runcmd.bat file

[2] With graphics.
  - Recommended configuration: P500Mz and 128Mo
  - Take a look at the preferences.ini file
  - Use the runit.sh or runit.bat file

[3] Analyse of logs
  - Recommended configuration: P200Mz and 64Mo
  - Use the runviewer.sh or runviewer.bat file
  - Since Alpha 2 click on the Graph button from GUI


Read the lauch script file if you want to use specific options

*------*
| Tips |
*------*

- Using the 1.4 final J2SDK will speed up the application.
  As an example a 15000 turns, with 50 worms, 1 food, and 150x150 life room 
  - with 1.3.1-02 => 4 minutes
  - with 1.4 => 1m53 minutes
  This conclusion is true under Windows AND under linux. Thus, some bugs
  related to your X server may be corrected if you use the 1.4
- Under Linux, using the graphics can be very slow, but if you
minimize the visible life room it will increase the speed of the 
simulation.
- Activating the log of the DNA convergence degre is not a good idea
if you make a imulation with more than 100 worms. Since  this method
is in O(n).
- The log viewer is able to read any file tabuled, and with a string
looking like #<legend>X	Y1	Y2	Y3</legend>


*-------------------------------*
| I found a bug, what can I do? |
*-------------------------------*
Send me an email renaud91@hotmail.com

*-------------------------------*
|         Finally               |
*-------------------------------*
I am looking for a job, so if you're interested, 
feel free to contact me
at renaud91@hotmail.com
