Keyboard / Mouse Input Device Design??

I need to know how difficult it would be to design and build a device that is able to force hardware keystrokes and mouse operations into an Intel based motherboard. I was thinking that it could have one end plugged in to a serial port or USB, and the other end would be a Y connection to the keyboard and mouse ports. How difficult and costly would building such a device be?

Peter Olcott wrote:
  Quoted Text Here  
A keyboard wedge ? Similar to the smaller and cheaper barcode readers ?

snipped-for-privacy@SeeScreen.com says...
  Quoted Text Here  
It's a trivial exercise at the PS/2 port level. PS/2 keyboards are really dumb. USB is a lot more of a challenge.

says...
  Quoted Text Here  
So the Mouse PS/2 port would be easy, too? The device could use a regular serial port instead of a USB, would this make it easier?
  Quoted Text Here  

Peter Olcott wrote:
  Quoted Text Here  
See for yourself: http://www.hth.com/filelibrary/TXTFILES/keyboard.txt
What are you intending to accomplish?
-Le Chaud Lapin-

  Quoted Text Here  
I need to do the MOUSE and the KEYBOARD. I have patented technology that can intelligently recognize anything on any GUI screen. There are some applications that get their keyboard input directly from the hardware. There might be some applications that do the same thing for the mouse. I want a single solution that always works on the Intel platform that can be used for automated testing on any OS.

Peter Olcott wrote:
  Quoted Text Here  
Uh ?
Can you translate that into engineerese ?
As in "do the MOUSE and the KEYBOARD" !!!!!!
Graham

wrote
  Quoted Text Here  
I need a hardware device that can feed keystrokes to the keyboard port, and feed mouse actions to the mouse port using software to provide this data. This is to automatically test application programs as if it was a human operator typing with the keyboard, or operating the mouse.

Peter Olcott wrote:
  Quoted Text Here  
And how is it going to receive this data ? Any limitations ?
Graham

  Quoted Text Here  
Why does it have to be hardware? Since you already specified that you're running on Intel platforms, writing drivers for Windows will probably cover >95% of all your potential users, and if you toss in Linux and the Mac OS, it's probably >99%...

  Quoted Text Here  
I already got Windows 100% handled, I am looking into making this technology universal.

  Quoted Text Here  
If this is something for future PC's, bare in mind PS2 and serial ports will cease to be included on new systems. This is because Microsoft has told PC manufacturers to remove all the leagacy devices on the LPC bus (internal ISA bus), which are mainly the built-in PS2 mouse/keyboard, parallel printer port, joystick port, serial ports, floppy disk.

  Quoted Text Here  
Build it?
http://tinyurl.com/ye8rr3

  Quoted Text Here  
It must be able to create keystrokes AND every mouse action.
  Quoted Text Here  

  Quoted Text Here  
Conceptually, how?

On Tue, 31 Oct 2006 05:23:07 GMT, "Homer J Simpson"
  Quoted Text Here  
I suppose if you started with a ball mouse card and tricked up the pulses that normally come from the wheel photo cells you could make the cursor move around but I doubt you could actually hit the right spot in any repeatable manner. The buttons are just switches so they are easy. I found it was a lot easier to just use an application that could be operated by keystrokes.

  Quoted Text Here  
I am building a universal GUI testing system based on www.SeeScreen.com technology that can always intelligently "see" anything on the screen by recognizing graphical objects from their pixels.

Peter Olcott wrote:
  Quoted Text Here  
What's that got to do with a mouse and keyboard.
Getting an explanation of what you're on about is like getting blood out of a stone !
You want a 'box' that simulates keystrokes and mouse actions. That bit's easy. What's going to *drive* it and how is *that* going to communicate with the 'box' ?
Graham

Peter Olcott wrote:
  Quoted Text Here  
This makes things clearer.
The hardware devices you speak of could be prototyped in perhaps a week (for one device; the interfaces for mouse and keyboard are similar) and made reliable in perhaps a little over a month using a PIC.
http://en.wikipedia.org/wiki/PIC_microcontroller
This book will have all you need to know about the various keyboard/mouse interfaces. Note the $145 price tag (it's worth it):
(Amazon.com product link shortened)62310673/ref=sr_1_1/102-6543581-0078538?ie=UTF8&s=books
But if you don't want to buy the book, look at the PIC page + these links:
http://www.beyondlogic.org/keyboard/keybrd.htm http://www.computer-engineering.org/ps2protocol / http://www.electronic-engineering.ch/microchip/projects/keyboard/v1xx/keyboard_v1xx.html http://www.simandl.cz/stranky/elektro/keyboard/keyboard_a.htm
On your web site (www.seescreen.com) you mentioned your solution could be used for testing, but you did not say what would drive the devices to generate their "key strokes" and "mouse movements". Naturally, you would not want to hard-code the seqences in the device. The inflexiblity would severely limit its usefulness. Whether you hard-code or do a new burn per run, you could let the device be programmable via a serial port on a PC, and write software to do that. Or use a PIC with an RF interface (www.microchip.com). So you would not have to detach and retach the device for programmability. You would program via RF from a central machine.
One of 1000's of people who would be able to give you a more exact/reliable estimate: http://www.zebryk.com / [Disclaimer: I neither know, nor endorse this person.]
-Le Chaud Lapin-

  Quoted Text Here  
I need a way to feed keystrokes to applications that get their keystrokes directly form the hardware. I want the same sort of solution for the mouse.