[Council] VOTE: JEP-0011

Peter Millard me at pgmillard.com
Tue Apr 2 09:40:38 CST 2002


I'm going to vote a -1 on the current incarnation of this JEP. Here's why:

The main reason for this is that it doesn't seem that we are "allowing"
enough room for this protocol to expand. Specifically, I have issues w/
section 3 which describes the protocol for specifying the jid-types... It
seems that forcing all applications of Jabber to conform to these specified
categories is too restrictive and confining. Section 6 also states that
defining new categories has to be done with an addendum to the JEP or a new
JEP.

What I would propose is to have the protocol look something like (from
Example 2 in the JEP):

<iq type="result" from="jer at jabber.org" id="browse1">
    <item xmlns="jabber:iq:browse" category="user" jid="jer at jabber.org
name="jer">
        <item category="user" type="client" jid="jer at jabber.org/foxy"/>
        <item category="application" type="game"
jid="jer at jabber.org/chess"/>
        <item category="user" type="client" jid="jer at jabber.org/palm"/>
    </item>
</iq>

This allows all items to have a consistant schema and the schema can grow
and expand as we add more categories. I think we should define the
categories listed in the JEP as "standard" categories, and use the same
rules for expanded the type attribute for categories. (If I want to use a
non-standard category, prefix it with "x-").

This is the main issue that I have w/ the JEP.. I think browsing is super
important, and we NEED to get a standard going ASAP, but I also don't want
to implement a protocol that we aren't all comfortable with just because
it's been implemented in the server, and a few clients. (Winjab is the only
mainstream browser that I know of that really uses browse).

Hopefully we can get this fixed up and re-vote on it soon.

Other comments?

Peter M.





More information about the Council mailing list