Some of us have been chatting this up for a while, but the vision of AJAX front ends to SOA composite applications is being heralded with new vigor at this week's JavaOne conference.
Tony Baer at CBR (registration required) has a nice rundown, including:
JackBe, one of the tools providers, is a three-year-old company that just recruited three former members of Sun's SOA development team to spearhead product development, is looking beyond the JSF [Java Server Faces] with grander ambitions.
"When we saw their [JackBe's] solution, we saw an instant fit of Ajax and SOA," said [JackBe CTO John] Crupi, one of the Sun alums, who just joined the company a couple of months ago.
Although JackBe could not be specific about its roadmap, which is still under development, the direction is that Ajax can perform the final mashup of composite applications that are already aggregated by SOA.
When you combine the benefits of SOA with AJAX GUI advantages and the news on JVM-based clustering, we're getting quite close to the reality of scripting- and frameworks-oriented developers being in the drivers' seat when it comes to developing and deploying mission critical enterprise applications, composite applications, mashup portals, and large-scaling ecommerce types of applications.
This class of developer won't do it solo (yet), but the power of new armies of GUI-level developers being able to actively contribute in significant ways to what was hard-to-master, Java-wiring intense (long dev cycle) applications -- and put them into production swiftly and cheaply -- is a very big deal.
Oddly, at the same time as this trend-line of accelerating developer capability/productivity amid a larger global pool of developer talent, Gartner is saying development is dead. And Microsoft still thinks it can only be done within an all-in-one dev-and-deploy proprietary environment. What are those folks smoking?
Perhaps, inadvertently, the juxtaposition of the Gartner Symposium and JavaOne conference at Moscone this week have given us a polarity analogy into countervailing mega trends in IT. On one hand, SaaS is more viable that ever, especially for SMBs. At the same time: visually rich, data-laden, service-driven, process-agile, robust applications are to be more easily and cheaply created and deployed than ever -- both for internal as well as B2C commerce uses.
The same benevolant technologies will be used by both those buiding SaaS host operations as well as those banging out highly customized enterprise applications/Web services/SOAs. How to invest, then, likely all comes down to economics. But perhaps not in the way that the SaaS advocates see it.
For my money and competitive advantage, I would absolutely do both for a long time to come. That is, for any company over 50 people with some internal or contracted app dev talent, I'll use SaaS to the hilt while also doing as much as I possibly can to customize and deploy my own services -- mix and match. The adoption of SaaS is therefore accelerated by more on-premises and custom app dev, not the reverse.
No, application development isn't dead, it's going to be something more and more people do. Essentially anyone who can type and work a mouse will soon be doing application development. And that may be the biggest productivty boost since VisiCalc.