On 1 July 2011 15:32, Michael Van <[email protected]> wrote:
> +1 to Ioannis point. Perhaps I've drank too much of the OSGi kool-aid, but
> moving from OSGi to Tomcat seems like a retrograde movement for SMX.

Noone said moving from OSGi; just supporting either Tomcat or Karaf containers.


> OSGi
> solves quite a few problems that exist in Tomcat, especially from a
> classloader perspective. There are just too many advantages of OSGi over
> tomcat, and moving an application from OSGi to Tomcat to me is akin from
> trading in your Prius for a Gremlin.  Basically, its going backward in
> technology with no real reason.

I hear you. It depends on how you look at it though. If you're happy
with Karaf, stick with it. If you're a user who's using Tomcat, knows
Tomcat really well (like most Java developers), has never touched OSGi
and has a bunch of existing WARs that just work (and probably don't
work in OSGi as you're probably using a ton of non-bundles), then the
Tomcat option is very appealing. Sometimes simpler is better (as
there's less to go wrong & you spend less time fighting with OSGi
metadata) - sometimes you want & need the power of OSGi.

Each to their own though; the container mechanism for building class
loaders should hopefully be quite separate to how ServiceMix adds
value to projects like ActiveMQ, Camel, CXF etc.


> All that said, since the use of Camel/Karaf became mature, SMX's viability
> has been a question in my mind. Any change that shows the utility of SMX
> over or in conjunction with Camel/Karaf is welcome.
>
> Good luck!

Thanks! Its always worth remembering that ServiceMix kinda gave birth
to both Camel and Karaf :). Irrespective of what makes the class
loaders, I still think there's a ton of stuff that ServiceMix can do
to add value.

-- 
James
-------
FuseSource
Email: [email protected]
Web: http://fusesource.com
Twitter: jstrachan, fusenews
Blog: http://macstrac.blogspot.com/

Open Source Integration and Messaging

Reply via email to