Haralds proposal was designed to be orthogonal to the concepts we currently have and embed within existing protocols and API. The idea is to build something straightforward to solve real problems. 

A query language sounds nice, but I don't really see how it relates to problems described in the former proposal. If you want to discuss that idea further, I would suggest to move it to another thread.

On 31 Oct 2014, at 17:03, Stan Silvert <ssilvert@redhat.com> wrote:

Taken to its logical conclusion, this would become a full-fledged query language. 

You could just create a client-side library for it and use that as the "language" develops.  I think that would be a safe thing to do before committing it as part of the supported server side interface.

 How much of a problem would this be if it were not for GWT's callback-hell?  It's my understanding that GWT is fixing that?