 Good morning, it's a lovely August 18th here in the Pacific Northwest. My name is Chris and these are the top 5 open source dot com articles for this week. At number 5 we have Tiberius Heflin's incredibly useful article, How to Avoid Leaving Money on the Table with Salary Negotiation. Everyone looking for a raise or a new job should read this article. Too bad it wasn't around when I was 25. Besides basic good advice on the topic of compensation in the job interview, Tiberius offers some useful tools for deciding on what kind and level of compensation we really want and what we can afford to give up. At number 4 we have Scott Nesbitt's tips on how to write better error messages. Let's face it, we've all been victimized by what can only be described as user hostile error messages. Can we really think that we're writing good code if our problem reporting is poorly thought out? Scott clearly doesn't think so and I believe he's absolutely right. At number 3, Lewis Coles offers us some cool ideas on how to customize the browsing experience by using our own JavaScript code to make other people's websites work better for us. Lewis gives us an example of changing a website's thumbnail image size and also dips into making a simple chromium ad on using JavaScript. For me this article fit into that broad category of, hmm, never thought of doing that before, but now that I've seen it I'm inclined to give it a try. At number 2, Jim Hall writes GNOME at 20, four reasons it's still my favorite GUI. Besides the four reasons, Jim takes us on a trip down memory lane and it's nice to have the screenshots to show how things have changed since day one, and dare I say how much they have improved. Jim's four reasons have a kind of nice kaizen feeling to them, giving a sense of understanding of the flow of the environment and using that flow to maximize work effectiveness. Well done. And at number 1, Charity Majors explains testing and production, yes you can and should. No, she's not saying forget all that other pre-production testing, rather she makes the case that only in a production system are we going to see everything being exercised in a real world way. The corollary of this is that production systems need to be well instrumented in order to facilitate that production testing, and we need to take what we learned from that back into the development cycle. A great antidote to those still dreaming about the big deploy and go for coffee button. That's it for now. I hope you enjoy your weekend.