On Thursday 11 March 2004 13:41, J Aaron Farr wrote: > You mean extensions that need access to meta data? Okay. I can see some > of that now.
I assume you mean the meta-information. What I mean is that currently there is a "contract" in the source file of javadoc @tags. Stephen has the opinion that the Meta package can churn that into whatever the 'current flavour of choice' is. This means that there is no long-term commitment to a sustainable and evolvable runtime solution, which makes component specs 'shaky'. >From my perspective, the Component doesn't necessarily have any source code at all (generated from other tools than javac), so (if you like) the @tags are in a different problem domain from Component<->Container and Component<->Tools contracts. > But there's also the issue of an explicit object model versus something > more general like a context map which would make it easier to add new meta > 'tags' without having to adjust the API. I am not sure what you are hinting here, but I suspect it also lies in a contractual relationship outside the Component-Container interface, but I could be wrong. > But this would mean I would be taking on Stephen, which is a different > thread and can wait until tomorrow. :-) He is coming around to realize that there is not a Specification but many Specifications, each addressing completely different concerns. My ONLY concern for the moment revolves around "Portable Component" and a "Managable Component", and being managable outside the container in various tools. Contracts surrounding the construction of a component, assembly of containers, resolution of dependencies, aggregated components and all the other aspects are not on the agenda - - - yet! Cheers Niclas -- +---------//-------------------+ | http://www.bali.ac | | http://niclas.hedhman.org | +------//----------------------+ --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
