Stephen,

I was initially skeptical too, however I've been using it to write unit tests for Fulcrum's services lately. I must say its wonderful. A couple point about why I have been impressed with plexus:

1. The component lookup mechanisms. Plexus includes a mechanism to have multiple comonents under the same role by extending Avalon's ServiceManager instead of just:

ServiceManager.lookup( String Role );

there is:

ServiceBroker.lookup( String Role, String id );

and you can do things like load a component and pass a custom configuration to it which is awesome:

ServiceBroker.lookup( String Role, Configuration config );

Do I need to explain how useful these things are?

2. Test cases are extremely easy to write. The test cases automatically load the config file corresponding to your class name. And plexus has lots of test cases.

3. It is not mutually exclusive with other containers. That is, a component developed for ECM can work in plexus and vice versa.

4. I don't want to insult the avalon team, but it isn't vapourware. It is here and working very well.

I think you'll see if you start using it how different it is and why Jason started his own. I encourage everyone to take a look.

- Dan Diephouse

Stephen Haberman wrote:
My initial reaction is that I'd like to see Turbine stay away from
Plexus. Even if it a technically superior implementation of the
Avalon Framework, I think we should stick with the Avalon community
on this one as they seem to be moving towards the previously
mentioned SuperContainer for all of the right reasons (e.g. keeping
a solid user base and cutting down on fragmentation).

To me, it's comparable to the Torque/OJB situation. Why should
Torque keep maintaining all of its persistence code when OJB is
already devoted to it? Why should the Turbine/Fulcrum community
devote to maintaining our own (assuming Jason donates it) container
when Avalon is already devoted to it?

Though if Jason could give a few of examples of why Plexus is
undoubtedly the right container for the Turbine/Fulcrum community
over Avalon's Fortress/Merlin/SuperContainer roadmap, I could
concede that perhaps it's the right way to go (e.g. I vaguely
remember Jason mentioning something about dynamic loading...not
quite sure).

I've been really hyped about the Plexus container for a long time
now, but after lurking on avalon-dev for awhile, I've come to agree
with their SuperContainer approach and would rather see the work
that is being done on Plexus be funneled into Avalon's
Fortress/Merlin/SuperContainer effort.

Though maybe Plexus is the right way to go for now until the
SuperContainer is released... Jason, can you see Plexus (or the
features that you/Zenplex require of Plexus) being implemented in
Avalon's Fortress/Merlin/SuperContainer efforts in either the short
or long term?

Thanks,
Stephen


--
To unsubscribe, e-mail:   <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>

Reply via email to