The Feed Object Model

[James M Snell <ja...@gmail.com>]

The central and most critical component of Abdera is it's Feed Object
Model interfaces (all of the interfaces defined in the "core"
component).  Our lives will be much more difficult if these interfaces
are not stable.  I'd like to propose that for the core package only, we
adopt an approve-first approach to all check-in's following the initial
check-in.  In other words, any change to the core interfaces would need
to be proposed first on the dev list for a vote before the change is
actually made.  Doing so is purely for reasons of maintaining stability
and making sure that everyone is aware of changes being made to the core
API.

- James

Re: The Feed Object Model

[James M Snell <ja...@gmail.com>]

Works for me. The lazy consensus thing is really what I had in mind
anyway.  :-)

- James

Garrett Rooney wrote:
> On 6/9/06, James M Snell <ja...@gmail.com> wrote:
>> The central and most critical component of Abdera is it's Feed Object
>> Model interfaces (all of the interfaces defined in the "core"
>> component).  Our lives will be much more difficult if these interfaces
>> are not stable.  I'd like to propose that for the core package only, we
>> adopt an approve-first approach to all check-in's following the initial
>> check-in.  In other words, any change to the core interfaces would need
>> to be proposed first on the dev list for a vote before the change is
>> actually made.  Doing so is purely for reasons of maintaining stability
>> and making sure that everyone is aware of changes being made to the core
>> API.
> 
> I don't see any problem with asking people to propose changes to this
> stuff on the dev list before committing, but requiring an actual vote
> for any change to it seems kind of extreme.  Perhaps a lazy consensus
> would work.  If someone proposes a change on the list and nobody
> objects within a certain amount of time they can assume that it's ok
> to proceed.  I mean worst case we can always revert a change that
> turns out to be problematic, and avoiding unnecessary bureaucracy is
> generally a good thing IMO.
> 
> -garrett
> 

Re: The Feed Object Model

[Garrett Rooney <ro...@electricjellyfish.net>]

On 6/9/06, James M Snell <ja...@gmail.com> wrote:
> The central and most critical component of Abdera is it's Feed Object
> Model interfaces (all of the interfaces defined in the "core"
> component).  Our lives will be much more difficult if these interfaces
> are not stable.  I'd like to propose that for the core package only, we
> adopt an approve-first approach to all check-in's following the initial
> check-in.  In other words, any change to the core interfaces would need
> to be proposed first on the dev list for a vote before the change is
> actually made.  Doing so is purely for reasons of maintaining stability
> and making sure that everyone is aware of changes being made to the core
> API.

I don't see any problem with asking people to propose changes to this
stuff on the dev list before committing, but requiring an actual vote
for any change to it seems kind of extreme.  Perhaps a lazy consensus
would work.  If someone proposes a change on the list and nobody
objects within a certain amount of time they can assume that it's ok
to proceed.  I mean worst case we can always revert a change that
turns out to be problematic, and avoiding unnecessary bureaucracy is
generally a good thing IMO.

-garrett