On Wed, Jun 11, 2008 at 10:24 PM, Georg Ragaller <[EMAIL PROTECTED]> wrote:
> So far as I understood is this RDF stuff a kind of introspection tool
> which can be used at application runtime, i.e. the components have to be
> at least assembled (if not instantiated), unless it's possible to dig in
> to meta-information which is provided from the runtime. [That's my
> current picture of how it works, correct me if I'm totally wrong]
Partially incorrect. It built a complete application structure from
the ApplicationInstance, but that includes all the details of the
"static model", i.e. the classes involved.
> @Tag ( { "domain:financial" , "domain:tax", "country:DE" }
> interface TaxCalculationService {
> ...
> }
In principle, I like this a lot... But I think the above might not be
the best way to go about it. The reason I am concerned is that in the
above you will then have a dependency graph reaching out, and the
"search semantics" quickly blurs.
Perhaps an initial "tagging" of Modules would be the right level of granularity.
Perhaps it is a matter of two separate things;
1) Code "markup" of anything.
2) Domain "markup" of modules.
I am still a bit wary of doing bytecode inspection though, and would
prefer to go via the RDF introspection library.
Cheer
Niclas
_______________________________________________
qi4j-dev mailing list
[email protected]
http://lists.ops4j.org/mailman/listinfo/qi4j-dev
