chemistry-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jens Hübel <>
Subject RE: CMIS Browser Binding implementation
Date Fri, 15 Apr 2011 16:11:45 GMT
@Jeff: The short answer is:

Because the browser implies so many restrictions (same domain policy, just GET and POST, ...)
that from my initial hope to get a nice, slim and RESTful JSON protocol all what remains is
... well a browser binding. ;-) 

Browser bindings emphasizes the primary design goal to allow fully implementing a client within
a web browser (Javascript).

It is still a much better choice than AtomPub and of course suitable not only for browser
implementations of CMIS. As Florian pointed out it is less chatty, more performant and it
should be the preferred choice for all those environments where SOAP is not an option, not
liked or whatever. I really hope it will quickly be the de-facto standard binding.

BTW. all the implementations using the Chemistry Java code will more or less get for free
(once the implementation is completed).


-----Original Message-----
From: Jeff Potts [] 
Sent: Freitag, 15. April 2011 17:53
Subject: Re: CMIS Browser Binding implementation

This sounds very cool. One curiosity question: Why is it called a "browser binding"? If the
main differentiation between it and the other two bindings is that JSON is going back and
forth, that is obviously useful, but not limited to web browsers. Server-side JavaScript apps
could also take advantage of this binding, right? Or are there additional pieces to the binding
beyond JSON that only make sense in the context of the browser?

Just trying to get a sense of where this can/can't be used.


On Apr 15, 2011, at 10:35 AM, Florian Müller wrote:

> Hi,
> The CMIS browser binding proposal is about to be moved into the CMIS 1.1 specification
draft. The browser binding is a third binding beside the Web Services binding and AtomPub
binding. It is based on JSON and optimized for consumption by JavaScript scripts in a web
> We have a first OpenCMIS server implementation in a Chemistry sandbox. It covers about
2/3 of the proposal.
> Although the browser binding will only be specified for CMIS 1.1, it should also work
for CMIS 1.0. This binding is more efficient and simpler to consume than the other two. We
should foster its adoption.
> I would like to roll the sandbox project into the server framework and complete the implementation
there. We can turn it off by default. Whoever is interested can switch it on and play with
> Since this is major code shift, I would like to hear some opinions before I tackle this.

> Thanks,
> Florian

View raw message