dotnetguru.org has been sending a ton of referrers the last few days. Makes me wish I could speak French. Well, more than phrasebook french. .NET must be a huge hit among french speakers, considering that one tiny little link is driving all this traffic, above javablogs.com, which normally rules the roost.

How timely that Chris Sells mentions a .NET bugs registry. Yesterday, a co-worker tracked down what appears to be a bug in the .NET Framework. We've noticed for some time that our servers where we run a .NET Remoting server seem to be steadily leaking GC handles. Yesterday, Mark made the observation that even with a new customer coming online with a huge transaction volume (proportionally to what we'd been doing), the rate of the leak didn't increase; meaning that the leak wasn't related to how many transactions were being processed. This pointed towards something constant, in our case, the Cisco Content Switches that we use for load balancing. We were able to correlate the rate of the GC handle leak to the rate of increase in the TCPConnections Reset performance counter, it turns out that the content switch polls the port the server listens on by doing a TCP connect (3 way handshake) and then, instead of doing a normal close with a FIN packet, the content switch simply issues a RST to kill the socket. From the research we did, this appears to be legal, but .NET handles this badly, apparently by leaking 2 GC Handles each time a connection is dropped in this way (and possibly memory, though I don't have numbers handy).

To the extent spam can be made this generic and still work, there are still techniques we can use to recognize it. I can think of two ways to deal with an unknown domain. One would be to send out a crawler to look at the page you're being sent to. You could filter the page much as you would the body of an email. If widely used, this would have the amusing effect of pounding any site advertised in a spam with a large number of hits. [Paul Graham]

I was thinking about this idea a couple weeks ago, sort of a Vigilante Justice for spammers. Basically, not only would you filter spam, but you'd have a robot go back and start making requests for any URLs found in the spam. Paul's idea is more interesting in that it's a side effect of having a widely distributed system for verifying the content of emails. I discarded the idea because in spite of the annoyance of spam, intentionally flooding a server (essentially doing the same thing via HTTP that a spammer does via SMTP) isn't the way to be a good citizen. I've been trying out the statistical filtering present in Mozilla Mail, we'll see how it works. My ISP email account is so polluted these days that it should train pretty quickly. What I really wish for is something that plug into a mail program and do statistical classification on all my mail, not just junk. Now that I'm using Hep at home, I'm finding that I'd like to save and categorize some of the RSS items it serves up, and I could use something similar at work with Outlook.