Siliceous and Delicious

Snowpocalypse #3 in less than two weeks, including ridiculous travel delays (walked to work again this morning through half a foot of new snow). Where’s your global warming now, Gore?

Thought about the next-steps question some more and decided that while I won’t ignore the mathematical morphospace altogether I do want to start on the radiolarian lineage project. I know it’s one that’ll work, it’ll go in the thesis, and the sooner I get started, the better I’ll feel. And feeling good means being more productive, which means feeling better, which means being more productive, which means finishing before I’m 30.

This means tackling the question of what my work-flow and data management system will look like for the project. Dave’s advice on this still echoes clearly in the back of mind:

Do NOT try to do all the analysis in various Excel spreadsheets – that was ok as a first attempt in a very short time frame, but it is really very tedious, limited and error prone by comparison to a proper database like MySQL (for which there is plenty of decent GUI software) and/or stat language like R

With this in mind I browsed the library for a good SQL book this morning, but was a bit disappointed… Most of them seemed pitched at a level above my starting point, and the “for dummies” publication on SQL didn’t have much guidance on installing software, or using it, except for MS Access—which is not available for download and would a) cost lots extra, and b) require me to flip back and forth between Windows and OSX, which I’m keen to avoid.

One possibly enticing option that popped up on my Google research on the subject is a package in R called RSQLite, which seems to embed a full SQLite system within R. This would be particularly nice because a) I would like to continue to work in R, as that’s where I’ll hopefully be doing my actual analysis and plotting, and b) the thought of fucking around with complicated SQL installations (as, for example, with my last and entirely disastrous attempt to install SQL, whatever version that was… maybe mySQL?). The downside so far is that documentation on RSQLite seems to be a bit thin on the ground, but I’m hopeful that a bit more searching will throw up some useful sites.