Tricky randomization question

Hello,
I am designing a source monitoring experiment and am having trouble plotting out the best way to randomize the trials. The stimuli consist of 32 target items and 32 distractor items, which are matched. I want to present 32 of these items (randomly from target/distractor) without presenting the matched stimuli. Additionally, I want half of these 32 trials to be paired with source A, and half with source B (So, 16 A trials and 16 B trials). [I]Then[/I], I want to randomly present 8 A trials and 8 B trials mixed with 16 of the 32 items not presented in the first part of the experiment. It's quite a handful! So... in short, my questions are:
1) How do I yoke together the target/distractors such that a) one of the two items are presented and b) none are presented with its match? The yoking section of the manual is helpful, but I don't understand how to have DirectRT draw randomly from [I]either[/I] the target/distractor list, rather than presenting from only the first and yoking to the second.
2) How do I randomly select 8 of the 16 A/B trials to present for the second half the experiment? The manual does a good job of explaining "whatever x trial was, display that", but I don't know how to do "whatever these trials were, pick 8 of them and display them".
Any help would be much appreciated, thanks!
-Nick