cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stefano Mazzocchi <stef...@apache.org>
Subject Re: [RT] Cocoonlet
Date Thu, 28 Apr 2005 13:44:38 GMT
Daniel Fagerstrom wrote:
> Bertrand Delacretaz wrote:
> 
>> Le 27 avr. 05, à 16:41, Daniel Fagerstrom a écrit :
>>
>>> ...Then we need a name for "sitemap blocks". I propose to call them 
>>> "cocoonlets"...
>>
>>
>>
>> Frankly, I don't like the name - most of these "-let" names sound bad 
>> to me.
>>
>>> ....But I don't think the names "component block" and "sitemap block" 
>>> are any good...
>>
>>
>>
>> Dunno, I think they are fairly simple to explain:
>>
>> A Component Block contains dynamically loadable java components, meant 
>> to be used in java or flowscript code.
>>
>> A Sitemap Block contains dynamically loadable pipeline elements 
>> (Generators, Transformers, Serializers), meant to be used in sitemap 
>> pipelines.
> 
> 
> Might be but it misses the most important aspect of the sitemap blocks, 
> i.e. that they are a kind of light weight Cocoons that contain an own 
> (sub) webapp and that talk to each other through the block protocol. So 
> they primary role is as an active service rather than being a passive 
> package of components.

oh, c'mon, Daniel, this argument is completely bogus. There is no 
difference in 'activity' between a pipeline invoking a subpipeline using 
IoC and a java class calling another java class thur IoC. The 
distrincion between 'active' and 'passive' lost its meaning a long time 
ago in IoC-based systems.

I'm very -1 on cocoonlet, it stinks pretty bad.

The plan is to attack sitemap and component blocks in parallel, but this 
is a implementation solution, not a marketing one.

I'm sure at the end a block will contain both components and sitemaps.

You got what you wanted, now stop pushing.

-- 
Stefano.


Mime
View raw message