I tried out Smalltalk for a day and I was completely amazed by the some of the features that still haven't made it into other languages. One of the features that I remember in particular was the ability to search for functions by examples. I don't remember the exact syntax for it, but you could search by something like:<p><pre><code> "hi" => "HI"
</code></pre>
and Smalltalk would tell you "toUpperCase". Immediately after seeing this, I wanted the languages that I use everyday to be able to do it.
The article highlights several contributions of Smalltalk to CS, but here's a few more:<p>David Ungar and Randall Smith, originally at Xerox PARC, then Stanford, designed Self, a prototype-based dynamic JITted Smalltalk dialect. In 1991 they joined Sun Microsystems, where Self kept being developed.<p>In 1994, Urs Hölzle at Stanford wrote a better compiler for Self. He was approached by Dave Griswold, who wrote a paper in 1993 with Gilad Bracha titled 'Strongtalk: Typechecking Smalltalk in a Production Environment', and together with Lars Bak the four of them started a company to commercialize Strongtalk. They were acquihired by Sun in 1997.<p>Together, the Self and Strongtalk teams' work resulted in the HotSpot VM for Java in 1999.<p>- David Ungar went on to be a member of the 'Dynamic Optimization Group' at IBM Research.<p>- Randall Smith now works at the 'Modeling, Simulation, and Optimization Group' at Oracle Labs.<p>- Urs Hölzle was one of Google's first hires and its first VP of Engineering and significantly shaped the company.<p>- Lars Bak joined Google in 2004 and went on to design the V8 runtime and Dart.<p>- Gilad Bracha went on to co-author much of the Java Language Specification and the Java VM Specification, then started a company to design the language Newspeak, and now works at Google on Dart.<p>- Dave Griswold continued to develop Strongtalk well into the 2000s, but I can't find any current info on him. If you know what he's up to these days, let me know!
I'm not really an ST fan myself. I'm not averse to it, or anything, but it has a steep learning curve, because you have to learn the class library, the IDE, and the language, and the poor tutorials and documentation make it even worse. Lisp had way better docs, and could be written in any text editor, so that's what I learned.<p>However, if you want to learn Smalltalk, might I reccomend GNU smalltalk? it's got the excellent docs you'd expect from a GNU project, unix integration, and while it does have a class browser and an image, it supports plain old text files for code, so you don't have to master an IDE if you don't want to, plus it's good for scripting. However, it has (mostly) the same class library as Squeak and other smalltalks, so it lessens the learning curve somewhat.
I learned to program in Dolphin SmallTalk at a small business run out of barn next to a bunch of soybean farms. While probably not the best language to run a production application with (because of the tininess of the community), I agree that it's a great language to learn on.
I think the SmallTalk development environment is one of the closest things available to what Bret Victor was talking about in this article: <a href="http://worrydream.com/LearnableProgramming/" rel="nofollow">http://worrydream.com/LearnableProgramming/</a>. The IDE integration and live coding experience is fantastic. I really miss this feature in other languages/IDEs. It's also amazing how the refactoring engine in Dolphin is as good/better for a dynamically typed language than many IDEs for statically typed languages.
I remember learning smalltalk years ago and would have loved to use it commercially, it always had many problems though. Platform support, databases connectivity, performance, you couldn't make a shrink wrapped distributable. It was always stuck as an academic language imho, which is a shame. Even now lets say I wanted to use it to develop my mobile app, I don't think this is possible? Learn it by all means it has some great ideas.
Article very briefly mentions Squeak[0], which I think would be worth looking into if you wish to pursue Smalltalk as a modern language instead of simply a historical footnote. Alan Kay, Smalltalk's creator, continues to contribute to Squeak, which he considers closer to his original ideal of what the language should be than Smalltalk-80.<p>[0]<a href="https://en.wikipedia.org/wiki/Squeak" rel="nofollow">https://en.wikipedia.org/wiki/Squeak</a>
I think that's vacuously true in that learning any language that's dissimilar enough from the ones you already know can make you a better developer. (Smalltalk's nowhere close to the top of the "learn next" stack for me tho)