On Saturday 05 June 2004 05:31, Jörg Hohwiller wrote:

> seems that this list is more interested on philosophy about spam and the
> universe than disscussions about the philosophy of avalon - nice jokes
> about that spam anyways but I would like to see some response to my mail or
> is it misplaced here???

Not at all... Somehow, I have missed your mail, possibly due to it was sent in 
the middle of the transition from CVS to Subversion....
Also, postings that are very 'neutral' have a tendency of not getting 
responses, i.e. "ok, I read it, but nothing much to add or dispute..."

Anyway...

> On Friday 21 May 2004 17:12, Jörg Hohwiller wrote:

> > I can see there is commons-configuration (which seems as it came from
> > avalon) and commons-logging. I saw in the archive that there was a
> > discussion about logkit vs. commons-logging + log4j before.
> > I would like to know if you are planning to directly use c-configuration
> > and c-logging in the avalon framework API?

As for Merlin, the answer to this is "No, we are not.". 

In logging, I think we need a more powerful classloading mechanism than 
commons-logging can support, but I am not sure whether we can provide a 
plugin for commons-logging in Avalon Logging, but that seems to be "cake on 
cake".

I am not well versed in commons-configuration, and don't have any comments on 
why. It just hasn't been discussed.

> > I personally think the commons initiative is a great thing and I can
> > still remember myself digging deep in apache projects sources to find a
> > ready to use cache, pool or whatever implementation.
> > Now why not just using these APIs as part of the avalon API?
> > I can see that changing an API this way in general is a bad thing because
> > of loosing downward compatibility but on the other hand it is not a bid
> > thing to replace org.apache.avalon.framework.configuration to
> > org.apache.commons.configuration and org.apache.avalon.framework.logger
> > to org.apache.commons.logging (okay the last one is not just a string
> > replace).

Well, being a framework, we have a fair amount of compatibility concerns, and 
replacements are out of the questions. Either a compatibility wrapper is 
provided or both are supported, and the component implementor makes a choice.

> > So all I am suggesting is think about it.
> > Disscussions welcome...

Change for the sake of change is not something I would typically spend my time 
on, and if no arguments are put forward of WHY a change should be done, I am 
sure it won't happen. What does the change bring to us (except increased 
maintenance)?

Cheers
Niclas
-- 
   +------//-------------------+
  / http://www.bali.ac        /
 / http://niclas.hedhman.org / 
+------//-------------------+


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

Reply via email to