There’s been so much going on in the performance space lately that I’ve been snowed under. It’s difficult to know where to begin chronicling all of the progress. I’ll start with a few updates from Sitepen.

• Back in April, Kris Zyp had a great article for IBM developerWorks called Ajax performance analysis. The developerWorks crew puts out some great material, and this is no exception. Simply put, it’s one of the best articles I’ve seen on the topic, and it should be required reading for every Ajax developer. He discusses Firebug, YSlow, and some client-side instrumentation techniques.
• Old friend of the Perf, Tom Trenka, had a nice post about string operations across browsers in May. One of the more interesting takeaways is with regards to IE7 versus IE6. The net — there’s no longer any justification, if there ever was, for special casing string concat operations for IE.
• One of my favorite tools, Firebug Lite, has seen some “>dramatic improvements in the Dojo Toolkit version, as discussed by Mike Wilcox in early June. The features discussed: a popup mode that remembers size and position, ReCSS (so you can reload stylesheets without reloading the app), a DOM Inspector, an Object inspector, and a command line. They’ve definitely taken Firebug Lite a long way past the initial goal of offering a bare subset of Firebug functionality to IE developers.
• A few days ago, Mike posted another article — this time with a nice addition to the recent swell of client-side profiling articles. Mike whipped up a nice generic mechanism for tracker client side performance in a cookie to remove some of the tedium from generating a statistically relevant data set in your own browser.
• Finally, Alex Russell expands on the concept of lazy loading by creating a stub loader for Dojo. Weighing in at a slim 6kB (gzipped over the wire), this build of dojo.js is just the bootstrap code necessary for loading the main functionality, all of which is deferred until it’s actually called within an application. John Resig posted a follow-up regarding some of the clear downsides of this approach, such as the potential violation of user expectations.