Stephen McConnell wrote:
>
>
> Marc Schier wrote:
>
>> Hi Leo,
>>
>> we had discussions on this in one of the SEDA threads earlier.
>> http://marc.theaimsgroup.com/?t=103070906800001&r=1&w=2
>> http://marc.theaimsgroup.com/?t=103107642300006&r=1&w=2
>> http://marc.theaimsgroup.com/?t=103073443400001&r=1&w=2
>>
>> I have a component implementation that allows to declare methods or
>> interfaces to be handlers (instead of having a generic event handler
>> interface, which is not type safe) handling events generated from oether
>> components. Exceptions and return values can be handled as well, all
>> being
>> declared in the component's configuration. I agree it would be better to
>> have a container handle this invocation, but because of the
>> instability in
>> container development there was no way to merge these concepts yet
>> (This I
>> guess will change with the release of Fortress 1.0.
>>
> A smaill coment - nothing to do with SEDA but kicked off from your comment
> on a Fortress 1.0 release ...
>
> I believe that the release of Fortress will simply create (a) a comfort
> factor for the islands of ECM dependents, and (b) slow the progress on
> a consilidated containmentr architecture. Personally (ie. someone
> thinking differently), I would hold of from a release of Fortress until
> the meta solution was in place and a framework for migration from the ECM
> role manegement to meta was established and validated - without this
> Fortress is nothin more that an enahnced,, improved, more elegant ECM
> (appoliogies in advance to Berin if this if offensive - but I really
> belive that there is so much more that could achieved with a little more
> patience and continuation of convergence). With a release we will get
> bogged down in questions about compatibility ECM/Fortress/Phoenix and
> Merlin. I really think the responsible move is to think these things
> through - build
> on each other strengths, and come out of this with the total solution.
Steve, the ECM folks need a migration path. So far, Fortress 1.0 will
allow that migration path. What ECMers need is to first relieve
themselves from the dependence on marker interfaces. They are crutches,
and Fortress 1.0 will help in that move. Secondly, Fortress smooths
the transition from *needing* to write components using the
ServiceSelector, to being able to live with a default. That will help
components that want to transition from DataSource being accessed within
a ServiceSelector to it being accessed directly in most cases.
Once ECMers are used to those fundamental changes, and have a taste
for more power and flexibility, they can migrate to a Meta enabled
Fortress.
> What is good about the Avalon Framework is its simplicity and clarity - why
> ant we do the same thing on the Avalon container solution?
Because power and simplicity are sometimes opposing design forces. It
is difficult to find the right balance.
--
"They that give up essential liberty to obtain a little temporary safety
deserve neither liberty nor safety."
- Benjamin Franklin
--
To unsubscribe, e-mail: <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>
