I always wondered what was inside those things

Steve Mitchell's blog
To help you understand how the system behaves while your application runs, Leopard comes with DTrace, a comprehensive command-line monitoring utility that can probe almost any aspect of system performance. DTrace includes a large set of tools that cover most general aspects of system performance, including kernel and user space code. If there is an aspect of the system that you want to track that isn't already supported, you can create your own tool using the D programming language.Note that the built-in Ruby comes with DTrace probes. This means you can trace the execution of your app from Rails, all the way down the Kernel and back.
Someone / some people who have been responsible for integrating Ruby and Rails into the latest version of Mac OS X (Leopard) have written some notes on what was involved and how Ruby and Rails work in Leopard. The Ruby build is a customized 1.8.6 p36 and actually integrates into Xcode and Interface Builder.. now making it a relatively easy task to put together Ruby GUI apps on OS X. RubyGems is also installed with a smattering of the most popular gems preinstalled by default.The office is awaiting our shipment of Leopard. Out of the box perl, python, etc, is one of the reasons I fell in love with OSX. I built my ruby 1.8.5 from source ("how rustic" says Sean) a few months back. It should be neat to see examples of ruby gui apps on the Mac.
The quality of tools on the web is increasing far faster than most “boxed” enterprise software, and those web tools are free or low cost, and available immediately, as opposed to going through a much more involved procedure to get access to tools on the inside, or convince IT to make them available. Hence my two cables...And yes, I literally have two cables.
Back when the software development world was first trying to recreate the obsolescent mass-production culture of its day, the Japanese had already provided us with the answer. The ideal batch size is one.When I apply this concept to my situation two things pop into my head 1) We'd better be practicing continuous integration 2) we'd better have full test automation. Otherwise, we can't do one piece flow (or even few piece flow).