The next release of the Big Iron pre-alpha code is now available from the Files link of the Big Iron project page. New in this release is the full implementation of the ping, connection and chat protocols for the network layer with an improved GUI client test program. There have been no changes over the previous in the 3D engine or scripting layer.

Starting with this release, images, fonts and textures in the data directory will be released as a separate tar file. Download both bigiron-cvs20021019.tgz and bigiron-data-cvs20021019.tgz. Untar both archives into the same directory.

As with previous pre-alpha releases, do not try to do 'make install' - run the sample programs from the top directory (see the release notes for more information.)

Note: if you have previously downloaded the cvs20021009 release, you can simply copy or link the data directory from your previous install. The data files have not changed.

I've checked in some preliminary code for text chat. The protocol will support point-to-point, multicast and broadcast text messaging, originating from a client or from the server. The code is highly experimental, so if you are building the code directly from CVS, beware.

I will probably not be doing another CVS release until I finish implementing and testing chat (as well as a remote shell protocol I want to do for remote server administration.) After that release, I'm going to begin working on merging the terrain demo and networking code. The current mid-term goal is to basically get a VR IRC client/server where clients can move around a terrain and see each other.

I've tried to start building some developer documentation for Big Iron by writing up a description of the terrain engine used to generate the screen shots. You can get to the documentation from the documentation manager on the Big Iron SourceForge project page.

I'm currently working on the networking layer; I've checked in support for ping and connection protocols. I'm working on adding chat and a protected remote shell. A document describing the implementation will added shortly.

I've put out a new release with the latest CVS code, including the new terrain features (texture, coloring, fog, height-of-terrain), a scripting interface and the fundamental classes for the network layer. Check the Files link on the SourceForge project page. The release notes describe the test programs and how to run them.

Python 2.2 is now required by the configure script. The check works for Cygwin on my Win98 box, but I haven't tested Linux yet. Let me know if you have Python 2.2 installed and the script doesn't find it.

The CVS code has been updated with height-of-terrain checks (so the trees and the camera actually follow ground level now), vertex coloring and a simple noise base texture. I've also enabled OpenGL fog to give some atmospheric perspective. New screenshots are available.

The next major step in the terrain rendering code is detail texturing, but there are a lot of complicated issues to work out. I've got a rough idea of how I want to do the texturing, but my brain hit its linear algebra limit for this week. :) I'm going to lay off the graphical stuff for a bit and work on the scripting engine and network code.

After a break of over a year, we're getting development going again on Big Iron. I'm going to chunk the current CVS tree and start over. The current plan is to stick with plib - Steve's added some cool stuff since last I looked. I'm developing against plib 1.6.0 and libpng 1.2.4.

I've added WIN32 (using cygwin) as a supported platform. My current development platforms are 400mhz Celeron Win98 w/ GeForce4 and a 440mhz Celeron Linux w/ Voodoo III. I'm currently working on a view-dependant LOD terrain object based on Peter Lindstrom and Valerio Pascucci's algorithm. I've got most of the non-texturing related stuff working and it looks pretty good - I'll upload some screenshots later. Once I get the texturing basically working, I'll update CVS.

I've also redone the initial "iron" model with Milkshape3D (and done a better skin and walk cycle). For only $20US, this shareware 3D modeller-animator rocks! Screenshots will be updated shortly. I'll also upload an .avi of the walk cycle.

Well, a lot has happened this past week. Needless to say I've been a little distracted, but I am still working on the project. I was doing some profiling of gui-test and realized something that a lot of you C++ wizards probably already know; the STL is SLOW (at least too slow to be called in a game loop)

I'm going to have to redo my existing code and get rid of all the STL container classes and strings. I also realized I need to refactor some of the code to avoid uneccesary copying/constructing/destructing.

Blame it on Java - it's made me lazy. :(

I've added a GuiButton class (with prelighting, handling of mouse motion while pressed and appropriate event callbacks). Run gui-test to see a demonstration. The source snapshot has been refreshed with the latest code.

gui-test now shows off our first 2D widget; GuiLabel shows one or more lines of static text with various alignments. The test program utilizes a GuiLabel to show our first FPS counter.

I've decided to go ahead and do a pre-0.1 snapshot release for those of you who might want to take a look at the source without having to set up CVS. The source snapshot is available on the downloads page. It can safely be installed (it will go in /opt/bigiron by default) and can be removed by simply rm -rf'ing the directory.) I'll try to keep this file up to date with the latest "stable" code.

In my ongoing effort to check in partly completed subsystems, I've checkpointed my current work on the 2D GUI library. Here's a screenshot of the gui-test program, which illustrates some simple HTML table-like layout of GuiObjects and GuiContainers.

client-test and server-test have been updated to demonstrate a few features of the new ConnectionProtocol protocol class. The features implemented so far include optional password authentication, connection identification, timeouts, application-level callbacks for protocol events, player names and a few security ideas. /doc/networking.txt describes the protocol in psuedo code.

client-test and server-test now demonstrate a new protocol abstraction layer over the low level socket class. The test code uses PingProtocol, a simple UDP ping time measurement. The Protocol class will eventually have callbacks for various protocol-specific events, timer support for autonomous events, efficient high level message interpretation, support for multiple protocols on the same socket, etc.

I'ved added two preliminary classes to suport UDP client/server networking. There are two new executables (server-test and client-test) which exercise the new networking code. It's UDP-based and uses plib's low level netSocket class.

Also, the file manager is now fully functional. script-test can read and execute python scripts from search paths passed through the command line or specified in path.cfg.

I've checked in partial support for configuration files and a file manager that will keep track of global and local configuration files. The ConfigFile class supports "standard" property file format with '#' comments and the addition of shell-like variable expansion.

The code now also installs sanely (into /opt/bigiron by default - you can override this with configure --prefix=<path>) The test harnesses can be run from the installation directory.

There's now a Logger class in src/util that provides support for subsystem oriented logging with verbosity levels that are adjustable at runtime. compile-test and script-test both use this logger now. Use the -l option to adjust logging levels.

There is now some preliminary support for integration of Big Iron and Python. The Python interpreter is embedded through the ScriptEngine class. There's also a ScriptAPI class that provides callbacks from Python back into C++. Check out the script-test executable which exercises this interface.

Certain events beyond my control have kept me away from Big Iron over the last year. I'm just now beginning to get back to work on it. Since I'm starting again after such a long absence, I decided to clear out the developer pool, forums, web page and mailing lists. I'll be filling the web page out again as I get time and info becomes available.

I'm not going to really recruit developers again until I get some working prototype code into CVS. If you want to get involved anyway, feel free to drop me a line.

Unless someone steps forward to claim responsibility for a Windows port, I'm probably gonna drop it as a project goal. That doesn't mean I don't eventually want a Windows port, it just means the Linux version is the priority and I don't have time to screw around with inferior OS's. :)

Stay tuned to this channel for more information.

    -Kudzu