> I bet you were into the lovely EJB artifacts too?  Ha ha this is strangely
> reminiscent.
>

No, I wasn't.  Really, I don't see the correlation.  But, as you like RPC
abstractions and I don't, I think we can just agree to disagree on religious
principles, and leave it at that.  Maybe take it up somewhere else?


> Again, I don't think that making the developer more mindful is a good
> enough reason to make the API more complicated and client code harder to
> maintain.
>
> But if there are technical reasons why continuations generated by the
> compiler in JavaScript would not work.... well that is a different story!
>

Well, I thought Joel and Bruce covered those.  "Would not work" here extends
to incorporate "would be confusing or unpredictable."  My feeling as a GWT
developer is that overall GWT has been moving, from 1.6 thru 2.0 trunk, to a
position of being less confusing and more deterministic and predictable.  I
like this.

I think, boiled down, my feeling is that the impedance mismatch between
XmlHttpRequest and RPC continuations is just too steep to ever really
abstract away successfully, so I worry that any effort in that direction
would be bootless.  Mind, I would probably have said that several years ago
about Java-to-JS compilation in general, so if anyone can prove my belief
wrong, it's this gang here.

- R

--~--~---------~--~----~------------~-------~--~----~
http://groups.google.com/group/Google-Web-Toolkit-Contributors
-~----------~----~----~----~------~----~------~--~---

Reply via email to