A few years ago I was teaching a course in VBA to some biological researchers in New England. I had no idea at the time (and they didn't know how to explain clearly what they wanted) but they were really interested in bioinformatics. Anyhow, thier questions came under the guise of doing analysis with Microsoft Excel - counting up frequency of occurance in certain things, parsing and extracting data from cells. Not a real problem with just a little bit of VBA knowledge, but it would have helped a lot to know more what they wanted in the big picture.
We talked a lot about evolution and the whole science vs. religion tidbit at which point they introduced me to what was a key difference between scientific and religious epistemology: verification vs. falsification. Up to that point, even though I had gone to a very conservative university, I had not heard of such a distinction in how to know things (maybe because instead of studying logic, I thought I'd study ethics for my token philosophy credit).
How do we know what we know? That's an interesting question - especially as it pertains to understanding scientific discovery and religious truth. I don't want to be controversial, I'd rather just point out the existence of such tensions while looking at the mechanics of it all.
So, enter SQL Server. This is a database product from Microsoft that I'm teaching this week. I was reading Kalen Delany as supplemental material on SQL server's internals and I found a term that I didn't know: heuristics. Apparently the SQL server optimizer uses "heuristics" in its cost based optimization process vs. other systems that use "rule based" optimization (the order of tables you query, etc... ). For more clarity on this however, I learned that heuristics is quite a human concept. And very related to this whole idea of epistemology and how we know what we know.
This document explains some common issues related to heuristics in basic analysis.
11:02:36 AM 
|
