I think I understand what they do, but just to make sure that I'm on the right track, i'll illustrate to you my understanding of how they work:
1) I create the mulitply/divide node first using whatever means needed to generate it.
2) Then I change my operation to the appropriate operation I need applied to my input value.
3) I then use the connection editor to plug a certain attribute into the input1X value of the multiply/divide node (allowing the plugged in attribute to have the divide or multiply operation applied to it).
4) I then change the input2X value to have that be the value which the plugged in attribute is divided or multiplied by.
5) then I use the connection editor to plug the outputX value of the multiply/divide node into the appropriate attribute that I need driven by the above generated calculation.
I think this is how they work, if not please clear it up for me. The reason I bring all this up is because even when i follow these steps, my end result is wonky. I change the value of my driving attribute fully expecting the subsequent attributes to have the multiply/divide calulation applied to them, but what I get is crazy random value changes that don't correspond at all with with I'd expect the equation to output.
I hope this isn't too wordy, please help if you can. thanks.