Scott Hanselman's post on Spam drew a bunch of comments, mostly telling Scott that he should be using Outlook and SpamBayes. I'd like to point out that SpamBayes also has a POP3 filter, so you don't have to give up your favorite mailer. I should also point out that Mozilla Thunderbird has built in filtering, though truth be told, I don't think it's as good an implementation as SpamBayes, and Thunderbird has a few things that drive me nuts, like the lack support for sending plain text emails or using > to mark up original text in an email reply (Honestly, email programs are getting worse with time. I'm about ready to go luddite and switch back to elm). Spammers are attempting to poision their messages, but I find that a trained installation of SpamBayes isn't bothered too much by them; I see maybe one spam every few days. Anyway, Sam Gentile, posted a comment that caught my attention.
Everyone says it's "out of control" and evertone suggests some geek solution like the latest anti-spam program. None of these address the problems which are social and not technical. When people say "out of control" what are they are prepare to do? It is freaking way beyond out of control even with anti-spam filters and costing the economy billions. When do we band together and say enough is enough? When do we put pressure on our lawmakers to make this stop? When do we do more than say it's "out of control?"
I don't think that Bayesian filtering is any better than an avoidance strategy. The idea Paul Graham presented in the original Plan For Spam was to raise the cost of business for spammers to the point where the volume drops off because the easy money is gone. However, in order for Paul's plan to work, some of the big providers (Yahoo!, Hotmail, MSN, and AOL) have to get on board and provide filters to their users. I don't know about the others, but Yahoo! Mail's filtering is pretty lousy. However, Sam's suggestion to put pressure on lawmakers is doomed to failure on the face of it; there's simply way too many juridstictions. It's like trying to get rid of gambling or porn from the Internet, there will always be some country that's willing to host it in exchange for a cut of the profits. The problem is SMTP; it's way too easy to lie about your identity, and it costs the same to send a million emails as it does one.
Gary King has a post on the goodness of unit testing and Lisp. My latest read has been On Lisp and I've made one of the conclusion that Gary has: if you get why unit testing is good and why modules that testability and good design often go hand-in-hand, then you get why functional programming and side-effect free programs are good. However, Gary makes the point that Lisp is interactive and that it's too easy to test in the REPL. I agree that this is a trap, what I've found from working on WordUp! (the software that runs this weblog) is that without tests, I never get a good comfort level for what's working. I tend to work on WordUp! in fits and starts, and my tests, such as they are, give me a comfort level with code that I haven't worked on in a while and don't necessarily remember all the dark corners of that I wouldn't otherwise have. However, I do find that working in the REPL is much like TDD: I invoke a function and manually verify that it does what I thought. Putting those assertions into a test case is just good discipline. In a compiled language, you can do the same thing, but the penalty for doing exploratory tests that don't work is higher.
All Things Considered had a story on Independence Air, a former "feeder" carrier for United Airlines that has ditched UA for a go at being another low cost carrier. They seem to be serving the east coast markets, like JetBlue and AirTran. If Independence is successful, UA stands to lose some, though I still think that this spells further trouble for US Airways. After my last trip on US Airways, there's no love lost between me and them. Also, today Gary Potter had a post about Delta's high labor costs, relative to the rest of the industry, and Delta could definitely stand to lose here; Independence will have service from RDU and ATL starting tomorrow and CAE starting on June 23. Judging from a random check on their site, they seem to have a good number of departures from ATL, though service to northern cities seems to connect through IAD, but it's not like you have a choice of either the 6 AM or 8 PM flight. Now if only someone could do something to help out the Denver market...