[Framework-Team] Re: first comments on plip 148 (moving to CMF 2.1)

[Alec Mitchell]

On 9/13/06, Martin Aspeli <optilude at gmx.net> wrote:
> Indeed. My feeling is that GS has some evolution to do before it's
> truly a solid replacement for what we currently do (which grantely
> isn't so solid) - maybe we're just replacing one set of design
> problems with another; not because GS is badly designed, but because
> I'd argue that product install/uninstall isn't what it was designed to
> do. Still, having used GS on b-org and partially on Ploneboard, it's
> quite obviously a very useful way of doing installs (workflows, for
> example, are senseless to do in Python).

I would say the same is true of:

FTIs
actions
portal permissions
tool properties (portal_properties, portal_factory, ...)
resource registry config
default content
probably a bunch of other stuff that I can't think of off-hand

However, I think we certainly need a way to preserve AT's automatic
procedural FTI setup, probably for as long as AT remains a viable
framework for product development.  However, any helper methods to
allow doing so should be in AT, not in Plone IMO.  And no
monkey-patches should be involved if at all possible.

Alec