18 November 2006 by Tam.

Things got a little hectic this week - next week (in other words, starting now) is our fall break (also Thanksgiving). I can’t believe the semester is so far along.

Anyway, yesterday Olga and I really hammered out our object model. It was a lot like when we did our database model - a lot of very productive arguing at the kitchen table with pencils in hand. I am going to make a UML diagram of what we came up this week so we’ll have something for Dr. Paul when we meet.

Posted in Uncategorized | 1 Comment »

10 November 2006 by Tam.

Met with Dr. Paul this week.  We showed him our DB model and started talking about the object model.

For doing actual intelligence stuff, of course, we want to read stuff out of the database, put it into some type of objects, and do our work with the objects.  This is a big scary thing to think about (I get intimidated a bit easily sometimes), but we can try to start by designing the objects.

We also talked about needing to get some data into our system, and how we could write some software to put in fake data based on ranges & the distribution pattern of different items.  (For instance, if we expect patient weights with a certain mean and standard deviation, in a normal distribution, we can build that behavior into the data-maker-upper, and that would give us better sample data than just randomizing everything.)  Scary.

We also talked some about the need to do more research.  I want to start by re-reading some of the sources I used in our original grant proposal; I only skimmed them when we were writing that (in kind of a hurry, as I recall).

Posted in Uncategorized | 1 Comment »

5 November 2006 by Tam.

Olga and I met Saturday and we really did hammer out a database design.  (It was more that than a data model, but it should roughly suffice for both.)

One big question between us was how to view the different pieces of patient information.  Would we organize them around “visits” (like an office visit) or around individual items like tests, symptoms, etc.  One thing that’s kind of difficult to remember is that you can always make it LOOK like it’s organized any way you want - you can make the user interface look like it’s organized by visit if you want, but then take the data and do whatever you want with it, and later spit it out a whole different way.  Or the same way.  Or a third way.

Anyway, we had a lot of arguing and I think it was really productive.  (Of course by “arguing” I don’t mean that we were angry or anything like that.  But for the first time we had actual real arguments about what to do, and I thought they went well, and that’s important.)

I think our database model is going to work well, at least as a start.  It’s always possible we’ll scrap the whole thing later if it doesn’t work out.

Posted in Uncategorized | 1 Comment »