Today's screencast looks at a performance testing - when you test Smalltalk code for performance, you need to be aware of how the compiler operates - otherwise, your performance tests might give you results you'll misinterpret.
You can download the video directly here. If you like this kind of video, why not subscribe to "Smalltalk Daily"?
This question about an "Iron Smalltalk" (no momentum in a few years) made me consider the oft stated (and never really fulfilled) desire for "Smalltalk on the JVM", or "Smalltalk on .NET" - I came up with the following:
The reason these "Smalltalk for .NET" and "Smalltalk for the JVM" projects never seem to come off is simple - Smalltalk isn't just flat text in an editor. Smalltalk is the entire interactive environment. It would be fairly simple to get a syntax parser, but it wouldn't be Smalltalk. It would be Ruby or Python with Smalltalk syntax. Somewhat useful perhaps, but not really Smalltalk.
If you really built Smalltalk in one of these other systems, you would have to invest a pretty large effort - and then end up with a system that ran slower than any of the currently extant commercial Smalltalk systems. Given that, it's kind of hard to see the point.
This week's ESUG video is Jorge Silva's "How to be Rich with Smalltalk" (ppt) presentation - a talk on some work to integrate Smalltalk and Adobe Air (separate from the Glare code) that Jorge had been working on. To watch, click on the viewer below:
If you have trouble viewing that directly, you can click here to download the video directly. If you like this video, why not subscribe to Smalltalk Videos?