I’ve just finished reading the No Fluff, Just Stuff 2006 Anthology. Overall it’s a pretty good read and, being an anthology, I was able to skip a couple of chapters which I wasn’t interested in without losing the thread.
Chapter 15 by Glenn Vanderburg closes the book on a somewhat nostalgic note.
Glenn begins by harking back to the 80s when Bell Labs and others extolled the virtues of domain-specific languages - little languages that capture the essence of specific domains and so are highly-optimized for producing applications in those domains.
Glenn also detects a move towards dynamically-typed languages - observing that Ruby, Python and even Javascript (!) are becoming more popular as developers discover their increased productivity. Lisp and Smalltalk are also resurgent as a new generation discovers their benefits.
Glenn asks ‘Why now?’ - answering that Lisp and Smalltalk didn’t achieve widespread popularity first time around not because of any limitations they had but rather because of difficulty getting those languages to scale down to the feeble processors and memory limits available at the time. (Although it’s not impossible to produce small, fast Lisp applications with some work.)
Even though developer productivity was poor with languages such as C it was just efficient enough to squeeze out the more productive languages. Now that many applications have access to (comparatively) abundant resources the need to use statically-typed languages to squeeze the absolute maximum from your application isn’t so pressing and so more productive languages can be used.
(If you want to read what Glenn says in full the chapter is available online here.)