It’s the end of day two here at SpringOne Americas 2008, and it’s been a pretty informative day. Here’s the higlights of the sessions that I attended: “What’s New In Spring Framework 3.0 - The Next Generation” presented by Juergen Hoeller. - The coming Sprng Framework 3.0 release will drop backward compatibility for JDK 1.4 in favor of full support for Java 5 and Java 6, including full support for generics. - The concept of “Conversation Scope” that currently exists only in WebFlow will be generalized, and made available in Spring MVC. - There will be a new Spring EL, which will be compatible with Unified EL, but whic will also have advanced features. - 3.0 will provide first-class support for REST - Portlet 2.0 - Declarative Validation - Hibernate validators and JSR 303 validators will be supported in binding via annotations.

“Lessons Learned Applying Spring MVC” presented by Rossen Stoyanchev - A big change - the Spring Controller hierarchy will deprecated in favor of POJO controllers that use the @Controller annotation. - The intended new approach is “annotations only” (on xml) for mapping URLs to controllers and views. - Also featured will be mapping by convention - a path of /owner/show may be mapped to, with no confguration other than definition of a ControllerClassNameHandlerMapping.

“Enhancing Spring MVC Web Apps Progressively with Spring Javascript” presented by Jeremy Grelle - An abstraction for rich client (Javascript) development - currently only Dojo 1.2 supported - looks like JQuery will probably be next. - Recommended approach is “progressive enhancement via decoration” - My impression was that use of this for client-side validation could be useful (though use for only this might be a bit heavy-handed), but that once you start to really use Dojo widgets, you start to get very invested in Dojo/Javascript. I think that most Java developers would be better served by using something like GWT or Flex for RIA.

“Working With Spring Web Flow 2” presented by Kieth Donald, WebFlow Project Lead - I picked up a few pointers, but since I’ve been using WebFlow 2 almost since it was released, much of this was a rehash for me. What I picked up was: - Best Practice - view templates and other resources in same subdir as flow, one flow per dir

  • if a view state has no view explicity named, the state id will be used to try to match a view

Keynote - John Rymer - principle analyst for Forrester Research, specializing in application architecture - Sees developers driving towards “lean software”. - Server consolidation - now only 4 major vendors - Microsoft, Oracle, IBM, SAP account for 37% of the market - no one else has any significant share - a result of acquisitions. - This consolidation isn’t working out well for all of the customers. - Cost is a major factor (opportunity for Spring tcServer and dmServer) - many shops are paying for full commercial J2EE server, but only using servlets


“Lean Software is the antidote to bloated vendor’s products and applications”

“Lean Software is community-led, with vendor complicity”

Spring is a very large part of Forrester’s thinking about this [Lean Software] movement.

We see a swing away from the WS-* specifications, towards REST services, due to simplicity and flexibility.

The same thing is happening in the .NET community - towards things like Silverlight.

OSGi has arisen now - should be helpful by allowing greater modularity and composability.

Software as a Service, and Platform as a Service will become more widely used alternatives in the coming year, as companies concentrate on their core business, and delegate non-core IT functionality to the cloud.

Lean software shifts complexity, but doesn’t eliminate it - Spring as a framework gives developers better tools to manage complexity - but Spring (or any open-source software) alone can’t do it - smart developers who use Spring and OSS creatively are very important.

Lean is developer led - others still must be convinced that it will work (e.g. Operations staff).


  • Lean software specialists will thrive
  • SAP’s platform influence will shrink
  • Microsoft will beat IBM and Oracle to PaaS (Platform as a Service)
  • Sun shrinks by killing products, may not survive
  • Agile methods will be the norm
  • Cloud for commodity workloads (e.g. Dell runs their customer files on SalesForce)