Web APIs and Standards for Java
Our first session today was presented by Craig McClanahan, who as many of you know is the creator of Struts, JavaServer Faces, and (as I learned today) also wrote the original servlet container for Tomcat. The session was a little disappointing given the stature of the speaker and his depth of knowledge, mostly because of faulty expectations on my part for what the session would be covering.
I expected to see something of an enterprise perspective on how to develop in-house standards and APIs around the Java platform. Instead, McClanahan was presenting more of a historical perspective of his experiences with APIs and standards for Java itself. While this was interesting enough, it was definitely not what I had hoped for.
It was interesting to hear him talk about the ongoing debate over whether Java should be open-sourced apart from Sun's control. He discussed at great lengths the limits to which Sun is actually "in control" of Java, some of which I bought and some of which I didn't, especially given that he is a senior engineer at Sun. What's he going to say? In any case, it was interesting to hear a Sun perspective on the JCP.
One of the few really interesting or significantly relevant things I learned from his presentation was that as of JSP 2.0, you can configure the compiler to error out if it detects embedded Java scriptlets ( <% vScriptlet = "random Java code inside your JSP"; %> ) . He talked a lot about the virtues of separating business logic from markup/presentation. This segued a little into a conversation about how Struts and JSF specifically implement and encourage MVC principles.
He also touched on the standardization around a portlet API, which he compared directly with the evolutionary signficance of the servlet API. I would like to have heard more about this, but it sounds like he's saving the heavy ammunition for another session he's giving tomorrow about JSF. Eda and I are currently negotiating over which one of us will attend that one. It conflicts with one of the presentations we both really wanted to see (Migrating to EJB 3.0 - led by Oracle's EJB and TopLink gurus).
I expected to see something of an enterprise perspective on how to develop in-house standards and APIs around the Java platform. Instead, McClanahan was presenting more of a historical perspective of his experiences with APIs and standards for Java itself. While this was interesting enough, it was definitely not what I had hoped for.
It was interesting to hear him talk about the ongoing debate over whether Java should be open-sourced apart from Sun's control. He discussed at great lengths the limits to which Sun is actually "in control" of Java, some of which I bought and some of which I didn't, especially given that he is a senior engineer at Sun. What's he going to say? In any case, it was interesting to hear a Sun perspective on the JCP.
One of the few really interesting or significantly relevant things I learned from his presentation was that as of JSP 2.0, you can configure the compiler to error out if it detects embedded Java scriptlets ( <% vScriptlet = "random Java code inside your JSP"; %> ) . He talked a lot about the virtues of separating business logic from markup/presentation. This segued a little into a conversation about how Struts and JSF specifically implement and encourage MVC principles.
He also touched on the standardization around a portlet API, which he compared directly with the evolutionary signficance of the servlet API. I would like to have heard more about this, but it sounds like he's saving the heavy ammunition for another session he's giving tomorrow about JSF. Eda and I are currently negotiating over which one of us will attend that one. It conflicts with one of the presentations we both really wanted to see (Migrating to EJB 3.0 - led by Oracle's EJB and TopLink gurus).
posted by Anonymous at 1:21 PM
2 Comments:
No matter who attends the session on the Portlet API, I'd love to hear their feedback. Some of the things I've been hearing lately suggest that a lot of the applications we right here would be best served as portlets in a portal framework, and this may be an indication that the industry is leaning in that direction.
Regarding: "One of the few really interesting or significantly relevant things I learned from his presentation was that as of JSP 2.0, you can configure the compiler to error out if it detects embedded Java scriptlets..."That's very cool, and would be very useful in our dev/test environments, unfortunately OAS 10g only supports J2EE 1.3 (so JSP 1.2), maybe in OAS 11e or 12c.
Post a Comment
<< Home