I think LPCR's useful life has come to an end. The Raspberry Pi and Arduino are really coming into their own. I believe that using an Intel based platform with ATX power is probably not something you would design today. The LPCR has had a good run, and its been a fun project, but with all things its time to step back and re-evaluate the future.
When the LPCR project started, the motivation was straightforward: build a real robot — one capable of genuine mobile robotics tasks — without spending thousands of dollars on commercial platforms. The answer at the time was to take a recycled desktop PC and put wheels under it. A full x86 Linux system meant real multitasking, real networking, and any programming tool in the Linux ecosystem. No cross-compilers. No proprietary IDEs. Just Linux.
That was a reasonable answer in the early 2000s. In 2013, a Raspberry Pi Model B costs $35, runs the same Linux software, consumes about 5 watts instead of 150, fits in your palm, and can be powered from a small LiPo pack rather than a bulky ATX supply. The Arduino handles real-time I/O — motor PWM, encoder interrupts, sensor polling — with determinism that a general-purpose Linux system can never quite match without an RT kernel patch.
The case for continuing to develop the ATX-based LPCR as a recommended starting point for new builds is hard to make. However, the techniques developed for this project — PID motor control, PS/2 encoder hacking, message-based control architecture, timing compensation on non-RT systems — are all directly applicable to any robot platform, including Raspberry Pi + Arduino designs.
The code, articles, and documentation on this site remain relevant. The platform itself is showing its age.