Slashdot Mirror
Building Consoles For Fun
tierra writes "Indiviuals writing their own games is one thing, but try building your own console. Russ Christensen, and his team put together in class, dive into the fun of using an old Nintendo system to house their customized XSA-50 Board. They also uses a XSA Extender to hook their personal console up to a monitor instead of a TV. They programmed Tetris and Space Invaders for their console using a system they call CASM."
1) Unified memory architecture.
2) Close to Win32, but not quite.
3) Liberal use of hashing and checksumming.
The XBE executable format is actually pretty interesting - it's similar to Win32's PE, but with many more flags and tables - for example, bitfields for what mediums are acceptable to run the game off of (DVD, hard drive, CD-RW, etc.) Each section and the relocation address tables have a SHA-1 hash taken of it, the header containing the section offsets and hashes of each is itself hashed, digitally signed using PKE, and then encrypted.
The hard drive also uses the ATA spec's password protection, although that's already been bypassed and the drive dumped. In any case, most people with homebrew code are using the neXgen or EvoX dashboards to run an FTP server on the XBox for uploading and downloading files to the HDD, so you don't have to muck about with IDE cables.
Most of the modchips out there right now work by tying the chip enable pin on the on-board BIOS to to ground, and emulating the BIOS directly on the LPC bus to allow execution of unsigned code and ignore mismatched media flags. (BTW, kudos to MSFT for complicating things with a floating ground... more than a few early modchips were responsible for fried PCs while doing in-circuit programming.)
How exactly does CASM differ from C? I don't see any ASM type code in there, it just looks like C. I realize that it isn't quite C, but hmmm... I'm interested in more details about the compiler/processor.
Remember back in the day when Sun was trying to push their JavaStation box? It was suppose to run a JavaOS on an Intel processor. It wasn't all that successfull.
Well, the lab I used to work at let some of the undergraduates take home the stacks of spare JavaStations we had laying around. =) The theory was that if you could install Linux and Quake on the OS, you would have a fully functioning console box, open sourced, that could run Quake. For those of you who are interested in mass producing console boxes, here are the lessons learned:
1. You need to be able to store the console operating system on an image server.
2. You need a serial port connector, or a LapLink connector, or a BOOTP enabled network card in the console box.
3. You need to install your OS on a testing machine, and strip down all of the extra functionallity (notepad, emacs, vi, and everything else). This is the process of optimizing your open source operating system. Set up the console box as you would a normal gaming system.
4. Take an image of the machine with some product, such as Altiris LabExpert.
5. Push the basic OS onto new machines, as needed.
6. Write documentation, develop games, etc.
7. Get a CD burner and burn games onto the CDs.
8. Package machines and CDs (seperately, probably).
9. Sell machines at cost plus a markup for time and effort.
10. Attempt to do a few things well and specialize.
As I recall, the ColecoVision was made with completely off-the-shelf parts. Wouldn't be too difficult to re-build one of those things. It was probably one of the few video game consoles that you could actually rebuild from scratch.
I think. My memory is pretty hazy and I haven't openned mine up in years.
J
I love projects like this. May just the fact that they succeeded inspire others to try the same thing.
On a tangent, this is the kind of thing that's been very possible for some time, but most people blindly assume that it's much too difficult. Similarly, writing a compiler for a high level language is a relatively easy project. You could do it in a semester course, or a month of spare time, but mention "writing a compiler" to the great majority of programmers, even those with lots of experience, and they run away in terror.
Personal chip design reached critical mass back around 1994.