geronimo-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Vikram Goyal" <>
Subject Re: Servlet and JSP engine
Date Wed, 20 Aug 2003 07:31:08 GMT
Hi Greg and others,

My understanding is that the initial idea of Geronimo was to develop a world
class J2EE compatible container, separate from the exisitng Apache projects,
with two main goals:
  a.. integration of various existing and new code bases into a J2EE stack,
with those codebases existing both inside and outside of the project
  b.. certification of the J2EE stack
An easy interpretation of these goals (specifically the first one) is that
if an existing technology supports the components of the stack then it
should be integrated without having to develop is from scratch. This leads
me to believe that the sole aim of this project is to bring together
technologies rather than provide implementations, which I believe should
really have been the aim. If for example, we had an open source ASF licensed
EJB stack, we would have been left with providing an integrator to that
rather than building it from scratch, which we are doing now.

A little more understanding of the complexity of J2EE is required to fully
appreciate these goals. I think the scope specified in the project
documentation is closer to the truth than the goals:

a.. a complete J2EE certified server which is fully ASF/BSD licensed and
backed by a healthy open source community.
a.. to create a fully modular J2EE stack so that the Apache community can
use whichever parts of the J2EE stack they require separate from the J2EE
server project.

A <b>complete</b> J2EE certified server and a fully modular J2EE stack would
be incomplete without an implementation of the servlet engine. Even if an
existing servlet engine can be plugged in behind the scenes it is breaking
the notion of a complete J2EE certified server. If the end user can take the
servlet engine out (presumably, after all the servlet engine is modular and
pluggable), the J2EE certified server is NOT a J2EE server. I am not sure
what you would call it, a Web services engine?, an EJB container?

Besides, there are several complexities with the plugging in other servlet
containers. I am not sure how many of you have ever tried to integrate
Tomcat with Apache or tried to embed Tomcat as part of an application. The
embedding makes the control pass on to another applications space, in which
we have no control or say. Ideally that should be the case, after all that
is what OO is all about. However, an integrated engine, developed within the
confines of the complete J2EE engine has a better likelihood of being
robust, modular and performing better.


> Vikram,
> have a look at
> where I have described the approach that we are taking.
> The actual implementation of the web container will be tomcat and/or
> jetty - but I would like geronimo to wrap those implementations
> sufficiently so that you cannot tell which.   Thus for 99% of
> and management, it will be geronimo that you are dealing with.
> The Abstract Web container classes have been created, but are a
> little empty at the moment... we have been waiting for the
> model to stabalize.  Hopefully there will be some more progress this week.
> cheers
> Vikram Goyal wrote:
> > Hello all,
> >
> > I have been following the list with quite fascination since it began.
> > volume of traffic has been quite overwhelming to say the least.
> >
> > I noticed early on that a conscious decision has been made to avoid
> > developing a servlet and JSP engine and rather rely on embedding either
> > Tomcat or Jetty. Is this correct?
> >
> > If this is the case, is there a rationale decision for this? If we are
> > developing everything from scratch, then why not the servlet and JSP
> > as well? It makes more sense than embedding other implementations.
> >
> > Regards,
> > Vikram Goyal
> >
> >
> -- 
> /**************************
>   * Greg Wilkins
>   * Partner
>   * Core Developers Network
>   **************************/

View raw message