> -----Original Message----- > From: Jason Novotny [mailto:jdnovotny@;lbl.gov] > Sent: Thursday, October 31, 2002 11:09 PM > To: Jetspeed Developers List > Subject: Re: why do portlets inherit from servlets? > > > > That sounds fine in theory, but seems to make the implementation > kind of schizophrenic. So if portlets are servlets and they have an > entry in a web.xml file and a servlet mapping, then it would seem they > are really managed by the servlet container e.g. Tomcat. However > portlets don't implement servlets they extend them, so the only calls > that the servlet container can make are the init(ServletConfig), > service(HttpServletRequest, httpServletResponse) and destroy() methods. > How do the init(PortletConfig) initConcrete(PortletSettings) and doView, > doEdit, doHelp, etc, methods get called that are defined in the > WebSphere portlet API? If portlets are in fact their own objects than > they can be managed by a chief servlet otherwise it seems that the > portlet container is really a servlet container. Possibly IBM hacked > Tomcat as well to make their model fly. I'm beginning to think there is > really no good reason for portlets to subclass servlets. >
I don't think this is the idea. Nowhere did I read that the actual servlet container was to be responsible for managing portlets nor would they be registered via the web.xml file. I think the idea was to reuse common method signatures that applied both to portlets and servlets. Portlets themselves will be managed by a portlet container not by the servlet container. Scott
