So the ingredients we know about:
1) Pressing a button starts a program.
2) There are programs which can simulate a mouse click.
http://answers.yahoo.com/question/index?qid=20081009025642AAtcz0t
mentions this software...
http://www.autoitscript.com/site/autoit/
I have no experience with that software, so can't say
which procedures in the language would come in handy.
*******
To click a button in a window, you need the "window name" first, or
some identifier that accesses the correct program. The button
in the program dialog to click, is at particular coordinates.
You don't want to use absolute coordinates, relative to the entire
desktop, because you don't know where you've moved the window when
the scanner program is started. It could be "cascaded" with other
windows, and start at a different (X,Y) offset each time.
That's why, if the program's window can be identified by name,
it doesn't matter where on the screen it is sitting. If the
coordinate system is relative to the corner of the program
dialog window, the job might be a little easier.
If the scanning program's window was fancy, and everything scaled
when you moved the lower-right corner of the window, then
clicking the button under program control could be harder.
In which case, you'd check the documentation for the scripting
environment, and see if it can access the "names" of all the
widgets inside the window. If each button in the program
interface, has a "name", you can associate the fake mouse
click with the "named" button.
*******
Scanners equipped with an "auto-feeder", can automate the
scanning of stacked paper. That won't work for books though.
And cutting the spine off a library book, would not make
the crusty librarian very happy. My scanner has an optional
autofeeder, but I never purchased it.
I've seen one device, that holds a book open. And uses a
wand to flip the pages. The "scanner" in that case, is a
camera, rather than a flatbed scanner. That's an example of
a professional device used for book scanning. That might be
how a company like Google, would scan an entire library. Load
a book, and walk away, until the wand has flipped all the
pages for you.
Most people just "crush" the book flat on their scanner,
and scan that way. That only works well on CCD scanners
(good depth of field) and not on CMOS scanners (poor
depth of field). My scanner uses a CCD, and I can see text close
to the spine of the book. The difference is illustrated here.
My scanner would look similar to the top photo. There is still
a little area of the spine, that is out of focus, but I can
see quite a lot of it.
http://www.carlmcmillan.com/DOF/DepthofField.htm
*******
If you get the AutoIt to work, post back with the details.
I like seeing details on any hack you come up with. For
the next person that needs the help.
Paul