Wednesday, March 02, 2005

Bottom-up Datalog
I spent some time today reading chapter four from Bob's book, Deductive Databases and their Applications. Bob recommended it to me, as it touches on some ideas in set-based evaluation.

The first half of this chapter all makes intuitive sense to me, as it describes the processes behind a top-down evaluation of a datalog query. Once it moved into bottom-up semi-naïve evaluation, it started looking familiar, most especially because it describes the fixed-point termination as the "intrinsic" delta being reduced to zero. However, it was when queue-based Prolog interpreters and breadth-first searching were introduced that I discovered that these describe the implementation techniques that I'm using for the rules engine.

Coincidentally, one of the motivations described for breadth-first searching is to avoid infinite loops, as this had occurred to me as well. However, the real reason was to avoid re-running deeper rules when higher placed rules in the tree generate new statements which result in the deep rules being needed again. Still, it is interesting seeing the same ideas come out in a text like this, as it demonstrates that my thinking has been clear on the subject.

I also performed an update on the ontology to handle the sub-constraints of a transitive constraint. This just matched the RDF I'd already written, so no updates were needed there. That let me just move on to encoding some more RDFS rules. These rules are not too hard, but it would take longer to explain what to do than to just do it. I don't need these rules encoded just yet, but I will need them shortly, so it doesn't hurt to work on them now. It's nice to work on something easy for a change, especially since the OWL rules are not as easy to encode (the current requirement that some rules have for sub-queries gives me the shivers).

I still have a little RDFS to implement, but I might have that done by tomorrow.

There was a lunchtime meeting run by Assoc. Prof. Janet Wiles for postgrads in ITEE in order to demonstrate a new Wiki the department has put up. The idea is to accumulate a body of knowledge on how best to perform research. Part of this is because ever supervisor has their own way of managing their students. It is also the case that every project can be different in unexpected ways, meaning that some students may not naturally encounter some information that would have been useful to them.

I was convinced to have a go, and have put in a few elements, including a definition under the title: "What is Science?" After finishing this last one, I realised that I had quite a bit of hubris presuming to tell the rest of the department what I think science is. :-) At least I've had some subsequent feedback to say that I didn't make a fool of myself.

During the meeting the word "Blog" was also mentioned. I pursued this, as I think it would be interesting to browse the blogs of others in the school. It could provide an insight on the processes that others follow, the progress they've made, and the directions they are considering in their research. I tried to put this forward, and I think that I may have convinced Janet to look into it.

Of course, in the middle of this discussion she asked to see a demonstration of my own blog. While I know that many other people can potentially read this blog at any time, I tend to ignore that, so I felt rather naked pulling it up in front of a room full of people. Fortunately the font was small for a projector. :-)

Time to get political... I was a little frustrated at the news today, as the next round of university reforms have been proposed.

The first item is that HECS is being authorised to go up again. This will never affect me, but it worries me about the future. Many young people I know are opting away from university to take on a trade. This is not because they are not suited for university, but because a degree will never make you a lot of money, and the debt incurred will wipe out any wage advantages the degree may provide.

A trade on the other hand costs nothing to get, and even pays a small amount while it is being obtained. All the tradesmen I know at my age make more money than I do. I don't worry about that, as I like what I do, but the notion of increasing fees on university payments seems wrong in that context. Americans may see this differently, as they all have to pay. However, we have a higher tax rate, and are supposed to get services commensurate with that.

The second item is an increase in full-fee places. I personally feel that this is a thin edge of a wedge. As more and more full-fee places are allowed, the government will have the credibility to reduce funding to universities (or maintain it while the system grows, which equates to the same thing) until almost all places are full-fee places. These things don't happen overnight, but the current trend looks to go this way.

The final item is a removal from the requirement that a university do both teaching an research. Australia does not have the money, nor the forward-thinking attitude that are both required to take an investment risk in research. This applies to both government and industry. The result will be that no university will be able to concentrate on research, and drop their (full-fee paying) students.

Conversely, since the money is starting to flow through from teaching, it will not be uncommon for a university to reduce, and perhaps drop, its research. I believe that this would be disastrous for a teaching institute, though it is hard to define in a 30 second sound bite. The best example I can think of is the level of knowledge and experience that lecturers gain by being a part of cutting edge research. As a student, there is something inspiring to realised that you are being taught theories that are still in the process of being developed, and that the body of knowledge you are learning is an evolving thing. Reducing lecturers to a purely teaching role reduces the quality of the degree. You might as well learn it all from books.

Of course, the other problem with dropping research from various institutions is that less research will get done. I don't believe that I need to explain what I see wrong with that picture.

OK, that is enough of a rant. Perhaps not all of these recommendations will be implemented (though they generally are). I can always hold out some hope that things will not go this way.

No comments: