Ramblomatic

Really, I just feel that I don't have that much to say. Important things have happened, though, so I suppose I should chronicle them as such.

Two weeks ago today, laurinskii and I were rear-ended by a going driving about 50mph. Totaled the Mustang, which wasn't fun. On the upside, I was able to buy (with the money from the insurance company) my first car in my own name - a 2001 Volvo S60. Pretty nice car, I might say. These are just things that happen, though.

What I really want to talk about is that (for the first time) I have an idea of what I would like to do for my master's thesis. Basically, reasoning among groups of defaults in answer set Prolog. My advisor has an idea of how to do it, but (and this is the exciting part) it has never been done before. It would essentially entail writing a new programming language that would consist of ASP with modules, so that preferences could be expressed between ASP modules (ergo, preferences among groups of defaults).

At this point it's a purely conceptual thing. I'll have to develop syntactical and semantic domains. But I want to post some code on here, so here's an example (extracted from my professor's notes on simple priorities between defaults):

entitled(X,1) :- orphan(X),
                 not ab(d_1(X)),
                 not -entitled(X,1).

entitled(X,0) :- child(X),
                 -dead(X),
                 not ab(d_2(X)),
                 not -entitled(X,0).

ab(d_2(X))    :- orphan(X).

An explanatory note from the book:

The first regulation says that orphans are entitled to assistance according to special government program 1, and the second says that all children who are not getting any special assistance are entitled to program 0.

The first two rules are standard representations of defaults. The last rule says that default d_2 is not applicable to orphans. Notice that if Joe is a child and it is not known whether he is an orphan or not, then Joe will receive benefits from program 0 but not from program 1.

So, reasoning among groups of defaults could (in all likelihood) be reduced to a similar problem. At this point I'm not really sure how, but time is definitely on my side.