After some much needed rest and catching up on schoolwork, I'm back to working on some of my own projects this weekend. At the advice of Potter, a CSH alumni, I went to Wal-Mart, bought a $10 UHF antenna and connected it to my MythTV box. So far, the results have been great! A quick channel scan was all I needed to get ten channels, double the number of channels I was getting with QAM256 over Time Warner Cable.

I've finished adding NNTP support to pynewspaper and am using it on a daily basis. The new version of the code is available from the repository. My list of features I gave in my Sept. 6th post are still in the works. Earlier today I started playing with Kismet and it's drone functionality on a WRT54GL for the purpose of sniffing out rogue access points. This is not an extremely useful feature in a dorm environment where the 2.4 GHz spectrum has more noise than the city zoo in the middle of an earthquake, but the ability to passively seek out malicious users will be a very effective skill in a corporate environment.

Most recently, I've been tweaking this blog's layout and updating a few things. I'll post again if anything significant is added.

I gave a seminar for CSH on the Python programming language earlier tonight. I've posted the presentation (html, swf, keynote) as well as an "advanced" example of a game written in pygame. Please email me if you've got any questions/comments on how this seminar could be improved.

I've updated pynewspaper with some new code to eliminate the need for the buildpaper.sh script. I didn't like the dependence on an external script to begin with and felt that the program could benefit from the advantages of threading on a per-page basis as well as per-feed. Simply untar the source and run "python newspaper.py feeds/" from within the source directory. All of the default feed pages in the feeds/ directory will be built. If you want to customize your feeds, create a new directory and place a text file with a list of URLs in it for each page you want to create. You will need to manually update header.res if you want the correct links to be reflected on the resultant pages. I'm not considering this a final release, but I feel that I'm getting pretty close to a 1.0 status application. I still need to solve some problems with columns disappearing and make configuration of the page header links automatic. Once these remaining bugs are worked out, 1.0 will exist. Features slated for 2.0:

• Proper length columns
• Web based configuration of feeds
• Packaging goodness
• Some cool feature using AJAX? Search?