February 4th, 2006

This one goes out to the robots


V7ndotcom elursrebmem




This post is not really meant for my friends. In fact its not even really meant for people (at least not directly). Its meant for the programs written by google which crawl the web. But it seems like a good excuse to do a technical posting for my friends.

Google (and all the other web search engines) have little programs that crawl the web. They work a lot like your webbrowser, except they don't bother showing the pages to a person. Instead these programs just record what text is on the page and what other pages are linked to. From this google decides how relevant a particular page is when you do a search.

If lots of folks link to a page when talking about cheese, then that page becomes more likely to be the top result when you search for cheese. Being in that top spot is very important, not just for you to find information, but for folks who are trying to sell a product online. Where will French cheese related pages rank vs California cheeses vs Wisconsin cheeses. If you're in the business of selling cheese it matters to you a lot.

So much so that there are companies called SEOs (Search Engine Optimizers) which will help ensure your page is at or near the top. Sometimes they advise simple changes to your page that help google to better understand what your page is talking about. Othertimes they play little games like creating lots of fake pages that all talk about the page they want to push higher. It creates an arms race of sorts. SEOs try to push their stuff to the top, while google tries to automatically recognize genuine interest, vs the techniques SEOs use to fake it.


SEOs, like all hackers can be over simplified into two groups. Whitehats and blackhats. This is a tradition that goes back at least as far as old western movies where you knew who were the good guys and who were the bad guys, based on the color of their hats. Blackhats aren't always hiding in the shadows. BMW was just banned from google search results in germany as a penalty for playing these little games.



SEOs are always trying to get better at what they do. One of they ways they do this is open contests. Pick a made up word, one that doesn't exist anywhere on the net. Then see who can make their page the top rated page for that word. As long as this is all done in the open I think it benefits everyone. Especially when it helps to reinforce the idea that google is hard to trick, and that real popularity is hard to fake.

For the last of these contests that I know about, it was won by someone who wasn't an SEO. It was won by a guy who'd been writing a journal online for a while. He encouraged his readers to link to his page to send a message about SEO, and the idea that real popularity is earned not bought. And the message was fairly clear.

Now there is a new SEO contest. Top prize is $7k. Thats intended to encourage folks to really try. For an SEO that does particularly well this is great advertising. For many that will have to be reward enough. But there is an effort for folks like me, maybe like you, to use this as an opportunity to send a message or do some good. That random word at the top of this post helps the prize money go to Celiac research. I have one friend with celiac disease. Its hardly the most pressing issue on the planet, but its one more worthwhile cause. As I write this the charity page is number 6 on google's list. If you would like to help add this text to a blog entry or website of your own.


<a title="V7ndotcom elursrebmem for celiac charity"
href="http://www.watching-paint-dry.com/v7ndotcom-elursrebmem/">
V7ndotcom elursrebmem</a>





While I'm talking tech-smack I may as well throw out a few random bits.


Code Sermon is a podcast about programming. You can also just listen to the mp3s if you're not hip to the whole podcast thing. He does a 15 minute lesson on one small part of computer programming. Rarely does he offer any deep insight that couldn't be gained from 20 years of hands on experience. But a succint refresher is always nice. If you're new to code I strongly recommend it, as it might save you a decade or two. If a random episode catches you fancy, consider doing them all in order since he began the series with fundamentals.


MPIBlast - I've been accepted as a developer on an opensource project I use quite heavily. Its a tool for comparing pieces of DNA. Due to the amount of DNA being searched, this program is designed to work on dozens or thousands of computers at the same time. This adds a level of complexity that traditional programming rarely touches on. Also it is more performance critical than your standard windows application so I really get to flex my brain more than I would for a typical project. While I think the existing team is happy just to have some one else to do some of the work, for me its a big honor. I guess its a geeky version of being trusted with the keys to the car for the first time.

Old but extra geeky. This video is a great ad from IBM for Linux. It gives me a warm fuzzy that I find hard to explain.