[Foundation] thoughts from another commercial player

Matt Tucker matt at jivesoftware.com
Fri Jul 11 15:51:23 CDT 2003


Ben,

> Then take the example of WebDAV. It is a set of protocol extensions to
> HTTP that provide the ability to do web-based distributed authoring and
> versioning. In the same way, Jabber is a set of protocol extensions to
> XMPP that create, what you have called, a "full-featured IM system." Both
> HTTP and XMPP are complete and usable protocols without the addition of
> WebDAV and Jabber, and there is nothing preventing any users from taking
> them and using them as is.

I think WebDAV is actually a pretty bad example. Although it is built on 
top of HTTP, it's purpose is significantly different. On the other hand, 
XMPP and "Jabber" (as you're defining it) essentially do the exact same 
thing with a few minor feature differences. The criteria it makes sense 
to apply is: do the systems have the same purpose? If the answer is yes, 
then it makes most sense to have a unified naming approach. Since the 
protocol is already called XMPP in the IETF, we should stick with that.

> There is certainly nothing preventing folks who feel that they cannot
> promote the Jabber terminology from creating their out community, and
> their own set of enhancements to XMPP. Microsoft has often taken broadly
> used protocols, and extended them to serve their purposes better. I hope,
> however, that in the interest of broader compatibility across all
> "full-feature IM systems" based on XMPP, that those folks choose not to do
> so.

Yikes, this is exactly what we want to try to prevent. We think there is 
a real danger of this as long as the Jabber name is used, which is part 
of the reason we've made this proposal. Let's use vendor-neutral naming 
so that nobody is ever encouraged or forced to take XMPP protocol 
extension work outside of the JSF.

 > as Tony said, there has been no evidence presented that makes
 > me believe that I couldn't clearly communicate that what I do
 > is related to an open protocol, as opposed to a corporate entity
 > that does similar work with a similar name.

Our experience has been the exact same as Barry's. There is the implicit 
assumption that "Jabber" means Jabber Inc by our customers. As a more 
specific example, we have a product in development that is built on top 
of XMPP (much like WebDAV is built on top of HTTP). We made the mistake 
of using the Jabber terminology with a potential customer of the 
product. Even though Jabber Inc. has no products even remotely similar 
to this one, the reaction of the customer was "Hey, I just got a 
brochure from Jabber Inc. If it's 'Jabber' that you guys are doing, then 
why shouldn't we be using their server instead?" Of course, I was able 
to explain the difference, but I shouldn't have to. At that instant in 
time I knew we'd never use the Jabber name ever again to describe our 
products. :)

Jabber Inc.'s own homepage doesn't do much to help with the issue. 
Here's how they define Jabber at the top of their website:

"Jabber is a software platform for moving your ideas forward - in 
real-time. Deploy secure, scalable, business-class instant messaging. 
Extend presence into workflow, supply chain, and knowledge management 
systems. Integrate a real-time communications channel into existing web 
sites and customer service portals. And do so with ease using Jabber's 
open, XML-based, natively interoperable architecture."

If you dig into the site, you can find a page that further defines what 
Jabber is and it at least mentions that it's an open protocol (but 
without mention of XMPP) -- again, not much clearer.

How can this situation not harm the adoption and ubiquity of the 
protocol extensions we're developing in the JSF as long as those 
extensions are called Jabber?

Regards,
Matt




More information about the Members mailing list