-<p> I am considering removing Step 2, since <i>all</i> information needed for an AI to keep the board state is contained in the initial board state and then the movement result messages </p>
-
-<h3> End Game </h3>
-<h4> Query </h4>
-<p> At the end of the game, the manager program outputs the following: </p>
-<p> VICTORY </p>
-<p> To the winning AI program, and </p>
-<p> DEFEAT </p>
-<p> To the losing program. </p>
-<p> Both programs then have 2 seconds to exit succesfully, or the manager will kill them. </p>
-<h3> Invalid Responses and Timeouts </h3>
-<p> If any program fails to respond correctly, or gives an invalid move, or does not respond within 1 second, it will lose by default. </p>
-<p> In this case, the message </p>
-<p> ILLEGAL </p>
-<p> will be sent to the malfunctioning program, and </p>
-<p> DEFAULT </p>
-<p> to the other program </p>
-<p> Both programs then have 2 seconds to exit succesfully, or the manager will kill them. </p>
-
-<h2> Modifications/Clarifications to Rules </h2>
-<p> Refer to <a href="http://www.edcollins.com/stratego/"/>This site</a> for the original rules again </p>
-<p> Currently, the pieces taking part in the combat are not revealed; only the outcome of the combat is revealed. In a human game, the pieces would normally be revealed. I am planning to reveal the pieces, since not revealing pieces significantly reduces the value of low ranked pieces (normally used for working out enemy piece values). </p>
-<p> It is up to the AI program to keep track of pieces. The manager program will only reveal the identity of the AI program's own pieces; the other player's pieces will be marked with * or # characters. </p>
-<p> In a traditional game, two pieces of equal value will both be destroyed in combat. Currently, only one piece is destroyed and the outcome is randomly chosen. </p>
-
-<h2> Example Programs </h2>
-<h3>"Dummy" AI</h3>
-<p> Dummy randomly moves pieces. It attempts to not make illegal moves (although sometimes it still does). </p>
-<h3>"Forfax" AI</h3>
-<p> Forfax iterates through all possible moves and allocates a score between 0 and 1 to each move based on how desirable the move is. It then selects the highest scoring move. </p>
-<p> Forfax is pretty terrible, but can beat the Dummy AI <i>most</i> of the time. </p>
-<p> It is possible Forfax can be greatly improved by altering the scoring algorithms. </p>
-<p> Since we don't want the Sample AIs to beat every entrance, I'm not really going to try and improve Forfax unless I get bored. </p>
-<p> If you are writing an AI in C++, Forfax already contains a lot of useful functions. </p>
-
-<h2> Longterm Scoring </h2>
-<p> I haven't started a system for pairing AI's and keeping track of scores and winners etc yet </p>
-<p> I'll probably have a bash script that chooses who will play, and stores all the scores in files. I've done this before for a BrainF*** based "survival of the fittest" style thing, so I can probably recycle it </p>
-<p> </p>
-<p> I believe it will make things more interesting if programs are allowed to keep state of who they have played, and the various tactics used. </p>
-<p> Its kind of hard to implement this... at the moment programs are killed every time a game finishes</p>
-<p> Perhaps it will be easier if each program is allowed access to one directory, where it can create and read files? </p>
-<p> This would allow me to keep the manager program which actually plays the games seperate from the program for scoring and matching of opponents </p>
-<p> There'd probably be security issues with that though. </p>
+
+<h4> Screenshot </h4>
+<img border="0" src="screenshot.png" alt="Graphical output of 'stratego' manager program." title="Graphical output of 'stratego' manager program. Options '-g' for graphics and '-b' to hide Blue pieces that have not taken part in combat yet. Red and Blue are both linked to the 'asmodeus' AI. Taken with scrot on 7/12/11." width="327" height="344" />
+
+<h3> Protocol </h3>
+<p> For the sake of simplicity and keeping things in one place, the protocol is now entirely described in the <a href="doc/manager_manual.txt"/>manual page</a> of the manager program. All updates to the protocol will be reflected in that file. </p>
+
+<p> Major updates to the manager program or protocol will be accompanied by an email to the mailing list. However, it is probably a good idea to clone the git repository, and regularly pull from it. </p>
+
+<p> <b> Warning:</b> AI programs <b>must</b> unbuffer stdin and stdout themselves. This is explained in the manual page, but I figured no one would read it. It is fairly simple to unbuffer stdin/stdout in C/C++ and python, I have not investigated other languages yet. </p>
+
+<h2> Scoring and Results </h2>
+<p> The competition will be a round robin, with every AI playing three (3) games against each possible opponent. A points system is used to score each AI, 3 points for a Win, 2 for a Draw, 1 for a Loss or -1 for an Illegal response (counts as a Win for the opponent). Scores accumulate between rounds. </p>
+<p> The winning AI will be the AI with the highest score after all games have been played. In the event of a tied score, the two highest scoring AI's will play one more round consisting of three games. If the scores are still tied, the official outcome will be a Draw. </p>
+<p> When the competition officially runs, results will appear <a href="results"/>here</a>. There may (or may not) be test results there now. </p> <p>
+
+<h2> Sample AI Programs </h2>
+<p> Several sample AI programs are currently available. The sample programs can be downloaded from the <a href="http://git.ucc.asn.au/?p=progcomp2012.git;a=summary"/>git repository </a>
+<p> <b> Warning: </b> No guarantees are provided about the functioning of these programs. It is your own responsibility to ensure that your submission obeys the protocol correctly. If you have based your program off a sample, please double check that it obeys the protocol. </p>
+<h2> Submissions </h2>
+<p> You must submit the full source code, and build instructions or makefile(s) for your program. </p>
+<p> Also include the name of the executable or script, the name of the AI, your name, and optionally a description of your AI and its tactics. </p>
+<p> Please email matches@ if you have a submission. </p>
+<p> <b> Code which attempts to comprimise the host machine, or interfere either directly or indirectly with the functioning of other programs will be disqualified. </b> </p>
+
+<h2> Dates </h2>
+<p> The competition is now officially open. Submissions will be accepted until midday, Saturday the 10th of March, 2012. Results will be announced as soon as they are available (depending on the number of entries it may take several days to simulate the competition). </p>
+
+<h2> Clarifications </h2>
+<ul>
+ <li> We are using the newer rules described <a href="http://www.edcollins.com/stratego/stratego-rules-later.htm"/>here</a>. </li>
+ <li> Scouts may move multiple spaces and attack in the same turn (in some versions they cannot). </li>
+ <li> Bombs remain in place until destroyed by a Miner (some versions incorporate "single-use" Bombs). </li>
+ <li> Victory is possible by either capturing the enemy Flag, or destroying all mobile enemy pieces.
+ AI programs do not have to "surrender" (as stated in the rules) if they have no mobile pieces. The manager program should end the game (if it does not, please report the bug)! </li>
+ <li> The newest versions of the physical board game reverse the numbers of the ranks (10 is the Marshal, 1 is the Spy). Our system uses the original numbering (1 is the Marshal, 9 is the Scout and s is the Spy). </li>
+
+ <li> You must always make a move. </li>
+ <li> <b>Remember to unbuffer stdin/stdout!</b> </li>
+</ul>
+
+<h2> Questions? </h2>
+<p> <a href="faq.html"/>Frequently Asked Questions</a> </p>
+<p> Please email matches@ or post to #progcomp with any questions not on that page. </p>
+<p> <b>Last webpage update: 07/01/12</b></p>