Dhanji,
> At first blush I definitely think this needs to be addressed
> with some kind of attrition--whether that means removal of
> the spi entirely, I dont know. As per the proposal and some
> initial conversations with Paul, I understood that jsr311
> intends to create a *useable framework*, not simply a bunch
> of specified interfaces. In that light, I would be for some
> sort of spi. I think we'll get a better handle on this when
> we have some examples to play with.
I am all for a usable API. Usable in the sense that target applications
would *only* need to rely on this JSR API to be developed. Then these
applications should be easily deployable into containers such as JAX-WS,
Servlet or Restlet.
In the case of Restlet, I also want to let developers manually attach the
annotated POJOs within existing Restlet applications but this won't be
mandatory. I except to have some way to say take my annotated POJOs and
generate a deployable application (WAR) out of them.
> I very much agree. Also the title of the JSR is JAX-RS which
> implies "Java API for XML." Is there something specifically
> tying Restful services to xml?
> I think javax.net.rs or javax.web.rs is more appropriate, but
> am not particulary fussed about this issue--so long as it's
> not SAAJ 2.0 =)
Sorry I missed your reply before sending the other email about the recent
JSR title change. javax.web.rs is good too, javax.rws (RESTful Web Services)
could be another option more inline with javax.jws.
Best regards,
Jerome