If you want to make an attempt at hacking the code of the database activity, modifying the simple search would probably be easier than advanced search and you should find the relevant code in the middle of view.php. The sql of the simple search performs 'like' search on the string from the input field so you could try spliting that string to items (by some delimiter) and add a 'like' statement for each one of these items.
hth
Implemented in a new 1.9 plugin albeit not fully tested yet. See http://moodle.org/mod/data/view.php?d=13&rid=4925 and further info and status updates in MDL-26983.
See image for illustration of some search options (e.g. more than one criterion per search field, search any item from a list (IN)):
I'm not sure there is even this 1 difference in the computation. In both cases there are 2m permutations, and one of them, the empty (or unchecked) set, is already presented, so you can either submit the presented or try one of the other 2m-1.
There is a clear practical difference on the user side though. In the all-correct the user has to actively tick all the options for getting it right. There may also be a psychological difference but I don't know what it would amount to. I always prefer to require more than less work from the user (offering the user an opportunity to get creative by searching for shortcuts!). So if there is no psychological effect which cannot be adjusted by some training I would dismiss the none-correct type as an unnecessary complication.
Given a question with none of the options correct, is there a significant difference between this question and the question negated such that all the options become correct?
"none of these" rings a liar paradox bell. Should it be "none of the others", or "only me", or "pick me"?
I haven't yet seen a multiple question with 'all/none of the above/these' which was such that none of these options couldn't be dismissed as a lazy shortcut. But that's me and perhaps only me ...
Cheers