youki, on Thu Jan 5, 2012 2:41 AM, said:
If the RAM is in the cartrige i 'd prefer that instead of the Opcode's SGM way.
The RAM will be in the cartridge but the access to the RAM will be different compared to RAM connected to the expansion port, in terms of programming. There will be an extra step involved in reading and writing to the RAM based on having to pre-set the R/W signal going to the RAM by a read/write to a specific address. RAM connected to the expansion port wouldn't require this step. If your program requires a huge chunk of RAM where you'll be doing a bunch of reads at a time or a bunch of writes at a time it will be pretty fast. If you're doing a bunch of individual reads and writes and you need the fastest possible performance it could become a problem.
The concept should be similar to using the VRAM.
youki, on Thu Jan 5, 2012 2:41 AM, said:
But to be usefull, we would need emulator that support that and that we can access the ram easly via C not only asm..
Emulator support shouldn't be difficult. There are a few open source ColecoVision emulators out there. My first approach would be to have the original author mod the code but I could put something together based on an open source emulator if I had to.
What ColecoVision emulator are most people using for development?
EDIT: Forgot to answer this one.
youki, on Thu Jan 5, 2012 2:41 AM, said:
and that we can access the ram easly via C not only asm.
Bank switching and programmatically setting the RAM R/W inside the cart shouldn't be a problem in C but I don't have experience programming for the ColecoVision so someone feel free to chime in if I'm wrong. The most likely scenario would be inline assembly in the C code to read/write to the bank switch addresses but again, I don't have Colecovision programming experience so it might be just as easy to do it in C with an existing module.
Edited by -=Lindsey=-, Thu Jan 5, 2012 11:58 AM.