[Zope-CMF] Re: RFC: PAS and the (non?) future of members

2006-02-03 Thread Rob Miller
On Fri, 03 Feb 2006 07:04:30 +0100, robert rottermann wrote:
 rob,
 we German native speakers are used to word our opinions much more directly
 that Americans.
 I believe for a German speaker there is not a trace of insult in what jens
 expressed.

understood.  i'm sorry for overreacting.  i guess i was just a bit
offended that anyone would think i might want to see Plone's messy APIs
moved down the stack.  the whole reason i'm _doing_ all this work is to
clean things up; moving Plone-specific code down below is the last thing i
want to do.

-r


___
Zope-CMF maillist  -  Zope-CMF@lists.zope.org
http://mail.zope.org/mailman/listinfo/zope-cmf

See http://collector.zope.org/CMF for bug reports and feature requests


[Zope-CMF] Re: RFC: PAS and the (non?) future of members

2006-02-03 Thread Florent Guillaume

Rob Miller wrote:

i'm wondering if it's not time to rethink the entire idea of members as
they currently exist in CMF.  members were originally a necessary evil,
because the user folder implementation of users didn't allow for enough
flexibility to support CMF's needs.  now, however, PAS makes it possible
to encapsulate all of the necessary behaviour in the user objects
themselves, and it should be possible to eliminate the complexity of
wrapping the user object altogether.

over the last few days at the snow sprint here in austria i've been
working on a Plone-based product called Membrane.  Membrane implements PAS
plugins which allow portal content to be used as the authentication,
property, group, role, etc. providers for users.  it's quite nice, i
think, very flexible and powerful, and i think it contains ideas that
might do well in CMF itself.

even if the content-based plug-ins are not desireable, i think it's still
worth investigating the use of PAS and the idea of deprecating the
member/user duality altogether.  anyone else interested in this approach?


+1

CPS has been using using this route for a while and we never saw a need for 
differentiating between users/members at the object level (Member role can 
still be useful of course). In CPS, the memberdata/membership tools delegate 
most of their operations to the CPSUserFolder or the user objects themselves 
(which then often delegate operations to our directories but that's 
another story).


Florent

--
Florent Guillaume, Nuxeo (Paris, France)   Director of RD
+33 1 40 33 71 59   http://nuxeo.com   [EMAIL PROTECTED]
___
Zope-CMF maillist  -  Zope-CMF@lists.zope.org
http://mail.zope.org/mailman/listinfo/zope-cmf

See http://collector.zope.org/CMF for bug reports and feature requests


[Zope-CMF] Re: RFC: PAS and the (non?) future of members

2006-02-02 Thread Rob Miller
On Thu, 02 Feb 2006 18:14:15 +, Jens Vagelpohl wrote:
 On 2 Feb 2006, at 15:35, Rob Miller wrote:
 even if the content-based plug-ins are not desireable, i think it's
 still
 worth investigating the use of PAS and the idea of deprecating the
 member/user duality altogether.  anyone else interested in this
 approach?
 
 It does sound interesting, unless the user folder all of a sudden gets
 overloaded with all kinds of APIs that it doesn't need for normal Zope
 operation.

i don't understand what you mean by overloading the user folder w/ all
kinds of APIs.  i'm not proposing any additions to the user
folder at all. i'm talking about a custom user class, along w/ custom user
factory, authentication, and user property decoration plugins that move
the behaviour that is currently in the member object down into the user
object.  no additional features, less complexity, more flexibility. 

-r


___
Zope-CMF maillist  -  Zope-CMF@lists.zope.org
http://mail.zope.org/mailman/listinfo/zope-cmf

See http://collector.zope.org/CMF for bug reports and feature requests


Re: [Zope-CMF] Re: RFC: PAS and the (non?) future of members

2006-02-02 Thread Jens Vagelpohl


On 2 Feb 2006, at 23:01, Rob Miller wrote:


On Thu, 02 Feb 2006 18:14:15 +, Jens Vagelpohl wrote:

On 2 Feb 2006, at 15:35, Rob Miller wrote:

even if the content-based plug-ins are not desireable, i think it's
still
worth investigating the use of PAS and the idea of deprecating the
member/user duality altogether.  anyone else interested in this
approach?


It does sound interesting, unless the user folder all of a sudden  
gets
overloaded with all kinds of APIs that it doesn't need for normal  
Zope

operation.


i don't understand what you mean by overloading the user folder w/  
all

kinds of APIs.  i'm not proposing any additions to the user
folder at all. i'm talking about a custom user class, along w/  
custom user
factory, authentication, and user property decoration plugins that  
move
the behaviour that is currently in the member object down into the  
user

object.  no additional features, less complexity, more flexibility.


I mean sticking Plone-only stuff onto something that's a simple user  
folder and should be completely agnostic of how it is used.


jens

___
Zope-CMF maillist  -  Zope-CMF@lists.zope.org
http://mail.zope.org/mailman/listinfo/zope-cmf

See http://collector.zope.org/CMF for bug reports and feature requests