[Setup] Re: Permissions

Raphael Ritz r.ritz at biologie.hu-berlin.de
Tue Jan 3 11:24:37 UTC 2006


Philip Kilner wrote:

[..]

>>In those cases you may actually want to be able to define
>>permissions on instance base - as opposed to type base
>>which is what DCWorkflow does.
> 
> Does this refer to Plone content, or to other artefacts?

Well, whether you consider folders per se as content
or not can be a matter of long debates. But to answer
your specific question: if you consider an ATBTreeFolder
content, then yes, Plone content - realized via a copy
and rename of the 'Large Plone Folder' FTI and then
configuring the workflow tool to not consider this
type subject to *any* workflow (BTW: this is exactly
how CMFDefault handles folders in the first place; it
was a conscious decision in Plone to make folders
per se subject to workflow which - frankly speaking -
I've never understood. For almost all use cases I
came across it is more of a hinderance than a help
to have folders workflowed).

[..]
> 
> Whether working with current or future workflow, would not the
> "Plone-ish" way of doing this be to create a type with a "null" (e.g.
> one state, no transitions) workflow?

This would be one way; an alternative I've outlined above.

[..]
> I'm not 100% sure whether we are talking about Plone content here, but
> we all need to know our options (including the option to shoot ourselves
> in the foot!). Since I get so much leverage from using Zope products
> which "play nice" with Plone, I think that there is some benefit in
> making such options better understood.

Maybe the key thing to understand here is that Plone is an
application on top of Zope *and* CMF. Zope provides the
security mechanism in the first place. The CMF provides a
workflow tool which - among other things - extends Zope's
security mechnism to become *type and state* depended
(via 'portal_type' and 'review_state'). This is a *big*
improvement but still there might be cases where you don't
want or need that. That's all I'm saying.

Raphael

PS: If people think Plone content ought to have a
review_state I simply disagree.






More information about the Setup mailing list