Perhaps you could take a look at the driver development kit on the msdn site.

You could always try "usb programming C" as a search term...

I don't think the timing characteristics of USB are particularly reliable. If you don't need the machine to do anything else, it'd probably be easiest just to find a machine with a serial port and write to that from a DOS or Windows 95/98 (i.e. before Windows was properly guarding hardware resources) program.

Probably the safest thing to do for reliable timing would be to find a machine with an actual PC beeper inside and reroute the signal from that. The beeper signal can be programmed to connect directly to a hardware timer, so your code could just set the signal going then stop.

EDIT: I can't really post references now, I'll try to provide some later.

EDIT2: piccolo — you're not an American?

I hope the TV guy isn't seeing this.

Anyway, piccolo, you can always show us that device you created/were going to create that would produce infinite energy.

It would make you rich and end a lot of issues in this world. Please?

Just so you know, Ive bookmarked this thread, I'll be reminding you of this project for a long time now. So you better complete it.

Anyway, why don't you use inline assembler? I don't know what you are using, but in MinGW, you can do something like this:

// this asm code is actually bullshit. :(
void send_signal(int param)
{

    asm("mov %1, %eax");
    // more asm for sending signals etc.
}

You'll have to find some tutorials on ASM and sending data out through ports. Maybe this can somehow help you, though I'm not sure if its connected to what you want to do.

For creating a driver you will have to use MS Driver kit like Steve++ suggested.

I knew there was some way to do this. Basically, in Windows, you can open any port like it was a file and then send data to it and receive data. Take a look here.

If it doesn't have to be extremely fast, do it with djgpp and DosBox0.72.

I believe Piccolo has stated in the past that he was a U.S. citizen but has moved to another country (possibly on a shakedown test of the spaceship)

this is what i have to far guys.

I have taken about a serial mouse. I found that the the yellow and white wires go from 0 - 21.8 volts when i call

HANDLE SerialInit(char *ComPortName, int BaudRate);
in the program loop

the device only needs 5v so i put some resister in to low the voltage.

the new problem is that the voltage dose not go back down until the program is terminated. so the voltage is not pulsing like it needs to be.

so i am currenly trying to rig the SerialInit() so that it UN-Inits what ever it inits

and this will give me my pulse of 5 v

edit:
ok i got it to pulse by adding

CloseHandle(hCom);
at the end

but not the program is run a little slow

EDIT

THANKS A LOT GUYS
I got the proof of concept working.

Im going to call around to see about getting a pattern.

He won't need to file a patent. Once this time machine is working, he can just go back in time and get rid of the patent office.

yes it serial port. im using yellow and white wire from a mouse.

i am using this to do the pulse

hCom2 = CreateFile(ComPortName,
GENERIC_READ | GENERIC_WRITE,
0, // exclusive access
NULL, // no security
OPEN_EXISTING,
0, // no overlapped I/O
NULL); // null template

CloseHandle(hCom2);

im tuning it to get the best resalt with this set up the handle calls are pretty slow but i should able to make up for it

yeah thats what i wanted to do that but it would haven taken longer to get the prof of concept up and running.

the read and right where handing up the program when i tried them. probably because the protocol was a blocking one

do you know good any docs on how the whole thing work.

the next step is it improve what i have here. all so im going to start researching my real signal transition miedem

thanks that dose look faster ill give it a try.