If you look at the "Avalon" big picture you can cut it into 
"specification" and "implementation". The Specification 
correspond to the Avalon stack shown below on the left.  "Phoenix" 
packaged as a RI along the lines Leo suggested makes lots of sense to 
me - but the cornerstone server/demo composition needs a clean-up. 
Repackage cornerstone demos as "Excalibur" and you start to see a 
marketable structure.

Avalon
======

   Specifciations                 Implementation
   ==============                 ==============

   |------------------------|     |------------------------|
   |                        |     | Excalibur              |
   |                        |     |------------------------|
   | avalon.atlantis        | <-- | Phoenix                |
   |------------------------|     | RI implementation      |
   | avalon.camalot         |     |(including Cornerstone  |
   |------------------------|     | services)              |
   | avalon.component       |     |------------------------|
   |------------------------|
   | avalon.Configuration   |
   |------------------------|
   | Avalon                 |
   |------------------------|


I think that the introduction of something more stuctured will 
facilitate a broad level of participation in the process, leading
to higher recognition and use.

Cheers, Steve.


> >More on sub-projects
> >====================
> >Generally, no two projects should depend on one another
> >(while cornerstone can depend on the framework spec, it
> >cannot depend on the impl as the impl depends on
> >cornerstone).
> 
> unfortunately thats not really viable - unless we go through and create
> factories for everything. In some cases blocks require knowledge of
> implementation because they need to instantiate and manage
> context/CM/configuration objects. (ie see James block in James poject).
>
> >
> >Ideal setup:
> >
> >1 - Avalon specification 'working group'
> >2 - Avalon components 'working group' which creates usable
> >    chunks of code for use in implementations and applications
> >    using Avalon, probably according to demands from 3.
> >3 - Avalon reference implementation (phoenix) 'working
> >    group' which creates impls of versions of the
> >    spec from 1 and components from 2.
> >
> >    3.1 - responsible for the bulk of the RI
> >    3.2 - responsible for the atlantis part of the RI
> >    3.x - perhaps more sub-groups for parts of the RI,
> >          like the Logger stuff.
> >4 - Avalon documentation 'working group' in which all other
> >    groups also participate.
> >5 - Avalon deliverables/support 'working group' which creates
> >    distributions and other deliverables and helps people set
> >    those up, and probably manage the website.
> >
> >1 depends only on java in general
> >2 depends on 3
> >3 depends on 1 and 2
> >4 has no real dependencies
> >5 depends on 1 through 4.
> >
> >Branding would be the concern for 5. There's no problem with
> >both an "Avalon" and a "Phoenix" (sub-)brand, me thinks.
> >Perhaps number 2 deserves a nice mythological sub-brand as
> >well...


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

Reply via email to