The demo programs are desigend to be run from Shell. Anyway, thanks to the "CLI" tooltype you can also launch them from Workbench. To scroll you'll need a joystick connected to the second port. When holding down fire, scrolling will happen in 8 pixel steps, otherwise in 1 pixel steps. Clicking the left mouse button quits the programs. Mouse and joystick are accessed through direct hardware calls. Therefore one will have problems with certain mices/joysticks if they need special drivers or are connected for example to the serial port.

The following shell arguments can be specified:

SPEED Shows the speed of the scroll routines. The smaller the yellow area (Copper bar) the faster they are.
NTSC Uses the NTSC Norm for display instead of PAL. Do not specify this option if your monitor/TV does not support it! The scroll routines don't change their behaviour when activating this option. That is, the visible area will still be 320 x 256 pixels, but of course in NTSC mode vertically you won't be able to see all 256 pixels on your monitor/TV.
SKY Activates a copper sky. Works only with Scroller_XUnlimited, Scroller_XLimited and Scroller_XLimited_64.
FMODE <n> Allows you to specify the FETCH Mode, which is only possible with the AGA chipset! The FETCH Mode defines the kind of access to the CHIP RAM. The access speed and the possible screen modes depend on it. Possible values are:

0: 1x FETCH Mode (like ECS/OCS)
1: 2x FETCH Mode (normal CAS, 32 Bit Bus Width)
2: 2x FETCH Mode (double CAS, 16 Bit Bus Width)
3: 4x FETCH Mode (double CAS, 32 Bit Bus Width)

2 (double CAS, 16 Bit Bus Width) does not seem to work, at least not on my computer. No matter what values I use for DDFSTART/DDFSTOP etc., everything I get is a display where the first 16 pixel column is ok, the second is not (it is instead an exact copy of the first 16 pixel column), the third is ok again, the fourth is not (it is an exact copy of the third 16 pixel column), ... That's not so much of a problem, because there is still the other 2x FETCH Mode (and of course the 4x FETCH Mode), which by the way is also used by the operating system if a screen's bitmap width is a multiple of 32 (but not 64). I would be interested to know whether "fmode 2" on someone's computer out there works anyway. Let me know.
HOW This option is for showing how the scrolling algorithms work. A hires screen will be opened where you can see what happens to the bitmap internally when you scroll around.

Since the programs hack the hardware, disable interrupts, etc., there is a certain risk. When starting/quitting there might well be crashes. Therefore don't have important things running in the background. Instead do a save and quit these programs. The demo programs only do read accesses on the hard disk and use the operating system functions from dos.library for that. So it is very improbable that your hard disk or it's content gets destroyed and that similiar bad things happen. Of course I cannot give a 100 % guarantee.

Before starting any of the demo programs you'll eventually have to disable screen promoters and certain options in your RTG software. The programs only work correctly if AllocBitmap() and OpenScreen() return what they were requested to return. For example bitmaps in FAST RAM or changed DisplayIDs are not good at all.