Impressions from my experience attending The Server Side Symposium, where I spoke on Rich Internet Applications
I did not push any Macromedia products at all -- in fact, upon announcing at the outset that I would not attempt to convert this audience of hard-core J2EE architects and engineers into Flash designers, I received applause (ouch!) -- but judging from the questions I fielded after I spoke, I think the most successful part of my session at The Server Side Symposium was the demo I put together showing JavaServer Faces (JSF) producing a Flash-based Datagrid bound to a server-side JavaBean property. It was displayed on a JSP that also contained HTML widgets bound to the same server-side JavaBean instance. So the rich Flash component (a richer data grid that can be had with page-based table approaches, I daresay) was produced using custom tags, and shared the same Java-based data model as the JSP. I ran it on Tomcat, not JRun (nothing up my Macromedian sleeve), and took advantage of Flash's local data store to work offline as well as while connected.
Before that demo, I explained what an RIA is and how various server-side patterns apply to RIA development on the client (using DAO's for example, to access local client-side data storage). I realize I will need to follow up this blog entry with an explanation of what the heck a Rich Internet Application really is, yet I cleverly sidestep rewriting the definition for the moment and quickly shuffle along...
I also spent some time talking about SOA, and how rich clients in a service-oriented architecture essentially allow clients to behave as nodes in a workflow, or service and messaging endpoints, etc. In talking with some of those working with messaging and business processes, this struck a chord; in talking with most others, I got some skeptical and/or blank stares. Which is good to know. Personally, I think business process components and asynchronous SOAP messaging will make a bigger impact by far than the RPC SOAP stuff that's going on today in terms of web services, and it also stands a better chance of growing the J2EE platform than does other suggested new stuff. Only a minority opinion, your mileage likely varies.
I enjoyed the conference, but I was a bit surprised by a couple of negative things in the community. Some of the venom in the EJB and AOP conversations, for instance; on both sides, I picked up a good deal of over-statement and silver bullet talk, seasoned with testosterone-fueled ego and seemingly far removed from mainstream corporate development. I was also a little put off by by just how much certain very intelligent folks really do despise the JCP -- I have *plenty* of issues with it, too, despite sitting on the JCP EC and contributing to various expert groups, but I discovered that most of those who hate the process the most are those who never apply to participate at all, who don't have suggestions for updating it that can be acted upon, and who don't actually bother to learn what's going on in the JCP before telling others what they imagine is going on. This should probably not be surprising. Still, I think I did translate some of the fury and bile into reasonable suggestions that I can take back. The process does need improvement, no doubt, and criticism of the constructive kind is helpful.
I also left with a couple of things to learn more about: I have to try Hibernate, for instance, as I was really impressed by those folks. I also want to do some customer visits and learn if the Java web developer really does want the half dozen VS.NET clones that Java vendors are demoing these days, which we saw at JavaOne as well as at this Symposium. I'm not so sure, but then I don't build Java IDEs.
As usual, the best bits weren't the sessions, but the conversations. Any chance to chat with folks like John Crupi, Kyle Brown, etc. is a good thing.