[Standards-JIG] NEW: JEP-0128 (Service Discovery Extensions)

Peter Saint-Andre stpeter at jabber.org
Tue Mar 9 23:54:47 UTC 2004

On Sat, Mar 06, 2004 at 06:55:11PM +0100, Jacek Konieczny wrote:
> On Fri, Mar 05, 2004 at 04:52:34PM -0600, Peter Saint-Andre wrote:
> > As discussed recently on this list, sometimes it would be helpful to 
> > extend service discovery IQ results. This JEP describes a recommended
> > approach for doing so by means of jabber:x:data.
> > 
> > http://www.jabber.org/jeps/jep-0128.html
> Why did you change jabber:x:data protocol and semantics??
> Data form of type "result" was supposed to contain a set of results
> of a data query including field names (table heading) and a list of
> records (table data rows) containing fixed field number. This protocol
> was accepted as Final it cannot be just changed, because someone would
> like it better this way.
> Now you made a JEP using the same namespace i quite different protocol.
> This is really evil. If the protocol is going to be defined this way
> no one will trust JSF extensions as good protocol specification.
> If jabber:x:data is to be used it must be used as it is defined in
> JEP-4. If that jabber:x:data definition is not good, then a new
> protocol, with new namespace should be proposed.

First of all, calm down.

Second, JEP-0004 was one of the earliest JEPs and I hate to say this,
but it was not written in a very formal manner. More recent JEPs are
written in a way that formally defines the protocol, whereas JEP-0004 is
quite informal. I have been cleaning up the early JEPs over the last few
months so that they provide a more formal description of our protocols.
JEP-0004 will probably be the last JEP that I work on because it is the
least formal of the bunch, and editing it so that it defines the x:data
protocol in a formal way without changing the meaning will be a job that
taxes the editorial skills of the current JEP Editor.

Third, I have not willingly changed the syntax or semantics of the 
jabber:x:data namespace. I tried to use the protocol in the context
we've discussed here on the list without violating the schema or what I
take to be the intent of the authors. However, because the JEP is so
informal, it is not 100% clear to me what that intent is. I do not think
that a data form of type "result" MUST NOT be used for anything but a
set of results in the "table" format described in the JEP right now, but
we can't tell from JEP-0004 because it does not use standard conformance
language (MUST/SHOULD/MAY/etc.) and is not written in a formal manner
such as is familar from RFCs and more recent JEPs.



Peter Saint-Andre
Jabber Software Foundation

More information about the Standards mailing list