Bertrand Delacretaz wrote:
> Le 24 oct. 05, à 06:50, Niclas Hedhman a écrit :
> 
>> ....I would like to propose;
>>
>>  1. Make Apache Cocoon 3.0 into the slim down, no frills, no blocks, no
>>     nothing, thing that has been discussed recently...
> 
> 
> Agreed, and even start this with 2.2, IIUC as soon as mavenizaion is
> finished, blocks can start living their own life.
> 
>>  2. Create subprojects in Apache Cocoon for heavily used blocks, which
>> "every"
>>     user can't live without. Since blocks are bundles, something like the
>>     Oscar Bundle Repository would make it a piece of cake for users to
>> install
>>     them...
> 
> 
> Even without OSGI, moving (physically or otherwise) all blocks which do
> not have strong community support to a "contrib" directory as suggested
> before would help a lot already. People who rely on those blocks will
> then have to find a way to have them supported, but it would allow us to
> focus on a primary mission of delivering a consistent, tested and
> well-documented core.
> 
>>  3. Create a user friendly "block exchange" elsewhere (cocoon-dev), where
>>     developers can publish and categorize their creations even if they
>> are
>>     not Cocoon committers. A kind of "Wiki brought to Coding", which I
>> find
>>     useful.
>>     The 'deprecated' stuff is moved here, for the sake of
>> availability, and
>>     allowance to fix bugs et cetera. Rhymes well with Daniel's
>> assertion that
>>     it becomes the user's reponsibility to maintain. As well as,
>> "Don't kill
>>     old stuff." and "Respect the past" expressed by others...
> 
> 
> Sounds good, and as a first step this could be our own contrib directory.

No it wouldn't, because we wouldn't agree what should go there. My
suggestion is to, as you did with the samples, leave all blocks as they
are, and start _hightlighting_ the ones we consider to be best
practices. Then, after some extended period if time we may decide to
purge the ones that have not received any highlighting, but highlighting
core blocks works much better than deprecating old blocks that people
may be using.

(And by way of recommendation of this approach :-) this is the approach
that Buddhists throughout the centuries used to deal with teachings that
had grown stale. They didn't say "that is a bad teaching", they said,
"hey, we've got a higher teaching".)

So, basically, until we've got blocks functioning, and have had them so
for _some_ time, we should do nothing more than highlighting and marking
up with meta-data for our blocks. Our blocks system and block repository
is going to create a new organism (which, yes, could well want a contrib
group elsewhere such as Niclas suggested), but I want to allow for that
organism to grow, well, organically :-)

Regards, Upayavira

Reply via email to