Lab Notebook: July 18

I only just started rerunning auteur on the fish data, using only body size for now. Exploitation and extinction data will be run later, after I figure out how to use the new features in auteur.

Although the script that I used to perform the auteur analysis is only around 30 or 40 lines of R code, it took an hour for me to get oriented in the code. This is because we have several copies of the same script (with small modifications) scattered across various folders in Dropbox.

After working with git and GitHub for so long, I find that the revision control that Dropbox provisions you with is not granular enough for working with code. Dropbox only provides file-level revisions, while git allows you to see line-by-line differences -- along with a commit message that explains why those changes were made. Without that level of detail, going into an old codebase can be a bit like walking on eggshells. You don't want to break known (or assumed!) working code to make a small change, especially if you have no idea how the code works in the first place.

I think Dropbox is still good for hosting and sharing data, but using git to control versioning makes things a lot easier for coding.

In other news, I need to respond to Ryan's email about phylogenetic informativeness re: the goby data, and I also need to finish writing a draft of the fisheries paper to send to Mike. Also, Dan's paper got soundly rejected from PNAS so there is a lot of work on the horizon to fix the issues and resubmit it to a different journal.