Just for laughs...

this week at work one of our coworkers is on vacation. She doesn’t like country music so my other coworker has been playing country all week while she’s gone. Today a song came on and this is what I heard, “A girl trying to find herself a perfect man is like trying to find a lettuce.”. A lettuce?!! I was thinking, “Gosh, that sure is a redneck country song! That kind of song would make me not want to listen to country music!”. I do like country music, I grew up on it. So I decided to google it and apparently I’m not the only one who heard lettuce, somebody’s grandma heard that too. The real lyrics are, “A girl trying to find herself the perfect man is like trying to find Atlantis.” by Jamie O'Neal.

Speaking of rednecks, the show My Name Is Earl is one of my favorites. We just recently started watching it and can’t stop laughing.

That Patriot Act is in the news again. Remember the law nobody can actually read? Supposedly, a college student was trying for an internship at a government agency, and the government did a background check on him via the Patriot Act. Something about this story seems fishy.

[An unnamed student] had created his Facebook.com profile when he was 18. Now 20, he had accumulated a good amount of material—typical college musings and photos—that his friends might enjoy but others might view differently.

The son was beginning a search for an internship, so [his mother] asked him to consider limiting access to his profile to just his friends. ... He heeded his mother's advice and did so.

Shortly after, he got the call he had been waiting for. A state agency wanted to interview him for an internship. ...

But, during the interview, something he was not prepared for happened. The interviewer began asking specific questions about the content on his Facebook.com listing and the situation became very awkward and uncomfortable. ... The interviewer explained that as a state agency, recruiters accessed his Facebook account under the auspices of the Patriot Act.

So, we're supposed to believe that the people doing background checks for the government, trying to get information that used to be publicly accessible on the Internet, would rather use federal subpoenas than The Wayback Machine or Google cache? But, wait, there's more.

Fortunately the son had previous working relationships with a few members in the office and knew a staff member there. He was offered and accepted the internship.

So, on top of the government agency's strange background checking behavior, we're also supposed to believe that:

• the student took the job anyway because of cronyism (which makes the background check strange);


• the student then blabbed to a news site, which figures that there are so many college students getting crony government internships after strange background checks, that they published the story with those details intact; and,


• that the government agency, that's already shown it likes to look online for dirt on prospective employees, isn't going to connect the dots.

I'm not buying it. Excuse me while I don't believe this story about an unnamed student, and unnamed government agency, and some very irrational behavior.

Suddenly many Saturday morning cartoons make sense. And, since I'm currently living in the South, I may start working on an anvil-resistant umbrella. I'm lucky to have made it through the Fourth of July alive.

Just to show that I'm not all technical, triticale (who often posts recipes) has something of a write-up on eastern North Carolina barbecue sauce. That's the stuff they serve around here. Basically, they smoke/slow cook pork overnight, tear it into itty-bitty pieces, put it on a hamburger bun, and put this sauce on top. I think it's wonderful, and I consider myself a transplanted Californian because I don't know how to make good barbecue yet.

The interesting part is that Brazilian barbecue (churrasco) is, at it's best, beef grilled quickly and occasionally sprinkled with salt water. The Northeastern Brazilians like to add onions to the water, but the idea is that you get the taste of grilled beef and hardly anything else. North Carolina barbecue involves grilling the meat without anything and adding lots of flavor through the sauce. I can't say which approach I like better. I can say that grilled meat is one thing that will keep me from becoming a vegetarian. Biscuits and gravy are another.

And, no, I wasn't born in California, so "transplanted Californian" isn't completely accurate. I was born in Colorado, but I only lived in Colorado a total of three years. I lived in California for about ten years and it had a much bigger effect on me than Colorado did.

OK, my gorgeous wife suggested that my last couple of posts were a little more technical than they ought to have been. So, to explain some of the backstory:

A programmer named Paul Graham recently wrote A Plan For Spam, which describes his attempts to improve spam filtering. Graham gave his advanced statistical filter away, and suddenly lots of companies advertised that they had advanced statistical spam filters. It's a well-written essay (as is his essay Why Nerds are Unpopular [which I only read out of curiousity]).

Graham noticed that companies tried half-baked ideas to cut down on spam. Some companies blacklist email addresses that spammers used. Others get together for a brainstorming session and decide that emails with certain terms (Viagra?) won't get through. It turns out that the return address on an email is just as easy to lie about as the return address on a real envelope. And if "Viagra" won't get through a spam filter, "Via-Gra" may. Using statistics solves this, and A Plan For Spam explains why.

The strange part is that many companies still use the half-baked ideas instead of the freely available good stuff. It seems odd to me, but I guess it's possible that some people haven't read A Plan For Spam.

Of course, math isn't exactly my strong suit. That's partly where the CRM Discriminator comes in. Don't understand the math behind the statistics? Worried that you'll write the wrong equation? Use the CRM's built-in text classifiers (written by MIT professors), and don't worry about it.

The really cool thing is that you can classify better than "spam, not spam." You can have the CRM Discriminator (or any statistical filter) put emails (or any text) into several piles. For instance, IF I had my own company, it would probably be a technical company. And IF I wasn't the only guy handling emails, I could have CRM look at all email coming in to admin@maxlybbertscompany.com and classify it as a Bug Report, Administrative, Technical Question, Financial Data, etc. That's great.

But I'm sure I could do more. I could have CRM look at several webpages I visit during the day (or, probably, feeds from those web pages), and ask "does this article look more like articles that Max likes, or articles he doesn't like?" If I would likely enjoy the article, I'd get an alert (or a web page, or some other link) telling me so. That's what I was suggesting was pie in the sky (because of other technical quesitons, namely databses), until I found out somebody else is already doing it.

But I'm still trying to think of a good use for the CRM Distriminator. Obviously you shouldn't buy a saw and then say "because this saw is so cool, I think I'll cut a hole in this wall here." That's probably what I am doing. But I'm on the lookout for a place where I can use my new tool.