2004/04/30 WorldForge Weekly News by: alriddoch

This is the WorldForge Weekly News for the week ending Friday the 29th April, based on what I have been working on, activity I have seen on irc, and commits to cvs.

Its been a very busy few weeks, with lots of success to report. Simon, James, John and I spent the 20th and 21st of April at the Linux User & Developer Expo in London, UK where we had been generously given space on the .org stand. Over the course of the two days we showed off a demo of sear running with a new Mercator based terrain engine connecting to cyphesis running Acorn style rules. Getting the demo ready resulted in a frenzy of last minute hacking, the main features of which were the last minute addition of an excellent chicken model created by Damien, a mass of work getting sear ready by simon, and some fixes to cyphesis restoring the AI to the level of functionality we had 3 years ago.

The big changes to sear mostly involved the addition of new terrain and sky code. Using the Mercator library, and based on the terrain rendering code developed for apogee, this code is now working well in sear. The user interface was improved to make demoing easier, including support for joystick control mentioned in the previous WWN, and support for mouse control of the view port.

This was the first time the latest development version of cyphesis had been used in a demo. This version's main features are its new persistence back end and support for Mercator terrain. A complete new map has now been created in cyphesis for the next few demos, featuring a square area about 1/2km each side, surrounded on 2 sides by mountains, and on the other 2 sides by the sea. Before the demo a number of long term bugs were fixed, including the lack of working AI, and after Ludex some performance fixes were made.

In the week or so since Ludex, James and simon have focused lots of attention on getting sear debugged more thoroughly, and have done lots of work on making sear more platform independent, so it works more smoothly under MacOS and Windows.

James has developed Eris 1.3 to the point where it is ready for field testing, and I have ported apogee to the new API as a first attempt. So far many issues have been fixed, both in Eris and Atlas-C++ 0.6, but the code does not yet run as expected. The new Eris code should have significant advantages over the previous series in terms of performance and simplicity, in part thanks to the more powerful API and internal performance improvements of Atlas-C++ 0.6.

All developers working on code that uses Eris should be evaluating the new API to ensure it still does what they need. Anyone who does not use Eris, but does use Atlas-C++ should be looking at the latest Atlas-C++ API in CVS. Feedback on both APIs is welcomed.

Ron Steinke has been hard at work on uclient, mainly focusing on Mercator terrain integration.