On 8 Jan 2004, at 15:57, Sylvain Wallez wrote:
So the question is: aren't we moving towards a 3.0 release?
I think that keeping 2.2 as a target keeps people focusing on back-compatibility while saying "3.0" upfront might scare users and give a false sense of freedom to break things to developers.
Even if changes are rather huge between 2.x versions, this will just please people. We don't need the marketing of big numbers for our users to understand why upgrading is good for them.
I mean, give them blocks and they will change to 2.2 overnight... but only if their things will work just fine.
What Carsten is proposing impacts probably 0% of our user base. Not only it's deep and undocumented... but it's stuff that even hard core cocoon devs know and date to touch... in case a super-power-user has used it, he/she will write to us and we'll tell him/her what to do.
I think impacting 0% of our userbase cannot be considered 'breaking back compatibility".
Keep in mind that this list is actively watched: if we were to break compatibility with stuff that a lot of people depend on, we'll sure hear from them sooner rather than later.
So I think the way we are moving on is just the optimal one: incremental action, keep back compatibility at maximum, but cleanup where required and listen for complains from the balcony when doing so.
-- Stefano.
