The biggest trends can be the quietest. Driving all types of application development for some 15 years now is Web application development, which continues to gain momentum well after we first began curiously clicking on links on that new fangled World Wide Web from Dr. Tim Berners-Lee.
Let's look back. FedEx came up with the first poster-boy app for the Web -- that being its Web-based package tracker. Then, software architects spent the first year or so of enterprise Web development building CGI connections to relational data stores. Java came in and had a grand old time building out Web systems. XML came in and, as a Web-oriented data definer, it begat Web services. Microsoft came along to combine Java-style programming with XML and called it .NET. People began to wrapper legacy mainframes with XML Web services.
Google and Amazon arose, building HTTP-based companies that up-played REST and down-played relational. They grew to be multi-billion-dollar concerns, and showed the biggest enterprises that whole new business models rode on REST and HTTP. Spring arose to challenge Java EE as a simpler means for fast Web app building - developers didn't need an all purpose object framework when the app's requirements boiled down to 'get it on the web.'
Web services begat SOA, and, today, soft-wired services feed cell phones over, you guessed it, the World Wide Web. Cloud computing is poised as a next step beyond Web services, with Amazon and Google acting as the poster-boy for the new effort. If I had told you 15 years ago you might be shifting your BPM applications to the Web, you would have laughed, but, with the advent of the cloud, this is being considered seriously.
Things change, yet some things remain the same. Some developers still struggle to find the correct granularity for services. Working in a hurry, they let the code override the software model but don't track back to synch these up. There will always be individuals looking to apply the latest silver bullet - can you say 'REST' or 'cloud'? - to each and every problem encountered whether it is malleable or not. We move forward in fits and starts. It only looks like a smooth progression long after it happens.