Planet SoC

Google IO
I spent Wednesday and Thursday of last week at the Google I/O conference in San Francisco. I had a great time, and particularly enjoyed the “URLs are people too” presentation (all the videos should be available on the google code website soon).
Slicehost
I’m done with shared hosting. 1and1 is ok for simple websites, but they’ve been having an unacceptable amount of downtime recently (enough that I notice it and it bothers me). I’ve decided to get a slicehost account, even though it’s a bit more expensive. I’ll be moving this page over there as soon as I’ve got the webserver running and stable. Now I’m paying for 3 different webhosting accounts… I’m planning to drop the 1and1 hosting plan (but keep them as my domain registrar), as well as possibly dropping the dreamhost account. Dreamhost isn’t bad, but their servers are slow, and slicehost is better. I may keep dreamhost around for storing data on.
I’ve been fantastically pleased by my experience so far with slicehost - their service is exemplified by the signup experience: 2 emails, one to notify me of the billing, the other to tell my slice was ready. No bullshit, no overly long explanation emails of how to use ftp and ssh, simply an IP, a username, and a temporary password. These people assume you know what you’re doing, which is exactly what I want.
It’s actually very handy having a few extra domains around that aren’t in general use. I pointed one of the idle ones at my new server, and have been using the domain for access rather than having to remember and type out the IP address every time I need to connect. I may redirect it elsewhere later, but it’s been working well for now.
GSOC
My Google Summer of Code project is progressing. I decided to call it DBGraphNav, and registered a googlecode project. Now I’m in the process of writing out the design document. I’m not going for a completely formal document, I but need something that will guide me in programming each portion. It’s available on the wiki.
My most recent accomplishment has been figuring out how to allow users to flexibly specify how to retrieve interconnected data. Each node on my graph will have a specified type, with a user defined SQL query that retrieves it’s connected nodes. Then each of those will have it’s connected nodes queried, and so on. This should allow users to build very flexible queries that include various types of nodes (e.g. authors, papers, research groups).
One very nice thing about googlecode is that you can use subversion to check out the wiki files, and then edit them offline in your favorite text editor. It’s extremely nice to edit things in emacs rather than messing with an interactive web editor. It helps me stay focused.