incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stephen McConnell <>
Subject Re: forking for @author (was: Code ownership)
Date Thu, 07 Nov 2002 02:08:01 GMT

Greg Stein wrote:

>On Wed, Nov 06, 2002 at 07:43:03PM +0100, Stephen McConnell wrote:
>>B. W. Fitzpatrick wrote:
>>>Peter Donald <> writes:
>>>>The best I have seen so far is a committer forking a project so he
>>>>can add his name as an @author - would have been funny except that
>>>>it effectively blocked progress in that project.
>>>Are you saying that this actually happened?
>>I'm guessing here but I figure Pete is referring to the fork of a 
>>sub-sub project in Avalon called containerkit, and the resulting 
>>excalibur/meta project.  For information about the differences between 
>>the two packages there is detailed message in the August Avalon dev achives.
>In that note:
>    "The existence of the Avalon Meta Model was brought into existence as a
>     result of a necessary fork of containerkit."
>What was the "necessary fork" ? The email describes the differences between
>the two, but not the reason for the fork.

The containerkit project was initiated by Pete as part of work on a 
component meta model.  I collaborated with Pete on that project - 
testing, bug fixing, enhanced documentations and so forth.  In parrallel 
I was developing the Merlin 2 container (a smart component and service 
management platform). The objective was to leverage a common meta-model 
within Avalon.  During the containerkit collaboration it became apparent 
that there were two ways to approach the construction of meta 
information - one approach, favoured by Pete, clesely following the work 
on Phonix a app-server withing the Avalon project (basically a static 
model in which meta data is declared once in system configuration 
files), the other approach, favoured by myself, followed a more dynamic 
model that addressed my own requireements together with requirements 
from other Jakarta projects (meta data creation programatically). 
Actually, the divergence between dynamic and static approaches to 
component management preceed my involvemnt in Apache (about 2 years). 
Unfortunately, collaraboration between Pete and I on this particular 
initative didn't work.  Pete made it clear to me that the constraints he 
was under relative to the Mirodom, Phoneix and a number of other 
projects prevented him from completely engaging in a collaborative 
community driven process.  This position, reinforced by vetos from Pete 
on changes required to handle dynamic services management requirements 
resulted in my lauching of a fork of a small part of the original 
containerkit project.  That fork is the excalibur/meta project.  It is 
alive an well and is used today by the Excalibur/Fortress, 
Excalibur/Merlin, and Excalibur/xfc projects.

In addition to the internal Avalon usage, the excalibur meta system has 
been doployed succesfully across the following applications:

   *  a container navigation and management application
   *  a minimal PKI system
   *  a sophisticated B2B collaboration and community Platform
   *  a really interesting dymanic security framework
   *  an implemetation of the OMG Notification Service
   *  an implementationof the OMG property service
   *  the proposed OMG registration and discovery service
   *  a CORBA ORB implementation
   *  an implementation of the final OMG PSS specification
   *  an implemetation of the OMG Interoperable Naming Service
   *  a distributed time service implememtation
   *  a domain management system
   *  a specialized CORBA service activation solution (that's makes
      CORBA really simple to use)

Cheers, Steve.


Stephen J. McConnell

digital products for a global economy

View raw message