views:

23

answers:

1

I am designing an interactive installation for a gallery where I will receive input telling me which of 8 input transducers have been bridged. For example if someone touches strip number 1, I will be able to detect that. For convenience let's notate that as {1}. If they touch 1 and 2 simultaniously, I will be able to detect that connection, let's call that {1-2}. If one person touches strips one and 2, and another touches strips 3 and five, I can detect the state {1-2, 3-5}.

In these lists of connections, any overlap between the sets will just create a union of the sets ie. {1-2, 2-3} would never be detectable, instead I would see {1-2-3}.

My job is to write code that makes events happen in response to these conditions. I will be polling the input and getting a list of groups of touched strips and then ...

So my questions are - what are the interesting properties of lists of subsets like this? What kind of patterns can I watch for? What is the formula for enumerating the list of possible groups of connections? The richer my insight into the properties of this data, the better I can map it to interesting and appropriate events. Mapping can be memoryless (ie. deterministic for a given input state) or it could reply to sequences, or even the timing of sequences. I have a few clues of directions I could take this, but I am hoping some folks with a bit more knowledge of algorithms and sequences will be able to give me some pointers here.

+2  A: 

You seem to be partitioning a set of transducers. Just consider any unbridged transducer a singleton set.

Doug Currie
Thank you very much! This is exactly the kind of lead I am looking for, regarding the individual input states. I have a bit more of an idea of what to do with changes over time (though I would welcome any suggestions regarding that as well).
Justin Smith
You're welcome. Re: changes, there are really only two things that can happen: two partitions join into one partition, or one partition splits into two. Lather, rinse, repeat. It is possible that these two basic operations happen so fast, relative to your sampling rate, that you don't see every step, but you can model it that way.
Doug Currie