Yes, and it is common practice in the PC industry to 'ground' yourself on the PC chassis (make sure it is the chassis). This negates any charge (and voltage) between you and the device. This will offer better protection than say, grabbing a RAM module and having the discharge path between you and the device go through the pins of the RAM module (and potentially knock out some transistors along the way).
However, grounding oneself on the chassis will not prevent other sources of ESD that include:
1) Fields from clothing or other materials that generate static (and hold a static charge, wood, clothing, paper, most plastics).
2) Will not equalize the potential between the PC and ground or you and ground.
So the best way would be to wear a wrist strap connected to ground, and place the device on an ESD mat that is grounded thus keeping all objects at a 0V potential. And also keeping the potential between objects at 0V (0V-0V = 0V) (and if your really cool you get an esd lab coat that shields your clothes)
A better way is ground oneself on the chassis (and keep contact through the duration of the repair) OR ground a wrist strap to the chassis (contact all the time)
And the worst way is to not ground yourself at all and have the grounding path go directly to a component on the PCB (via screwdriver) or through a component you are installing.
In ESD controlled environments there is much more to it then simply grounding yourself on the Chassis, but this is not possible for most people who just want to install a module.
You can't guarantee the polarity of the charge and so some form of rectification is needed to ensure that any current is passed through in the same direction irrespective of polarity. This could be achieved with two capacitors and two diodes so that you can store the energy from positively and negatively charged people on two separate capacitors.
The diodes should have an extremely low leakage to prevent charge being lost. The capacitor should also be chosen to have very low leakage for the same reasons.
I think I'd also consider having a spark-gap protection device across the input to the diodes to prevent over-voltage damage. If you use the carpet static value value below to work out the capacitance and voltage rating you should be OK: -
20 mJ = \$\dfrac{CV^2}{2}\$ and if you chose 1 uF, the voltage rating would need to be: -
\$\sqrt{\dfrac{2\times 20\times 10^{-3}}{1\times 10^{-6}}}\$ = 200 volts.
For 10 uF the voltage rating would be 63 volts
1 nF would require a voltage rating of over 6000 volts.
I was planning to have a metal plate on a wristband and the capacitor
connected to this and then to ground, this I think should charge the
capcitor from what I understand. Would this work?
I would consider the capacitor(s) and diode(s) to be fixed and let folk walking around discharge themselves to a plate connected via the diode(s) to the capacitor(s). It might be a good idea to issue conductive thimbles so that folk don't feel the spark that might jump to the plate. Safety should be considered and the diodes will prevent someone getting a full energy discharge if they touch the plate.
Best Answer
Sure. In your wimshurst machine example, you're already charging up two capacitors that are built into the machine, they're the jars on the right and left.
If you want to charge up another capacitor, you could connect it between the rod that's going into the jar and the outer bit of metal (though you'd probably need a capacitor for each jar), or charge it by putting it between the balls. However, you need to make sure the capacitor is rated to the voltage you're charging it to (on the order of 10kV), which is not common. Since you're making it yourself, you will need to use a thickish dielectric and be prepared for it to arc over. You also need to be prepared for it to take a long time to charge. According to wikipedia, a leyden jar has about 1nf of capacitance, so if you add say a 10nF capacitor to each jar, the sparks will occur about 10x less often.