[Foundation] Should the IETF standardize the protocol?

Adam Theo theo at theoretic.com
Wed Jul 31 13:25:47 CDT 2002


Reply inline.

Joe Hildebrand wrote:
> Theo said:
> 
>>I personally would like to find practical ways for the JSF 
>>[...] to retain full control over the protocol instead of the IETF, 
> 
> Sorry to pick out just one thing, and I know this is going to be part of
> the debate process, but I'm traveling during the 24 hours for debate,
> and I thought this was important enough to discuss anyway.

No problem, I'm glad you decided to make this an issue. I wasn't sure if 
I should open this can of worms now.

> I just want to make sure that people know that if you don't want the
> IETF to have change control over the protocol, it's the same thing as
> not wanting an IETF standard for the protocol.  The IETF demands change
> control as a condition of taking on the work.

If this is the case, then I still politely say no to the IETF. Besides, 
perhaps it is time for the IETF to change its ways a bit, and allow 
outside open standards foundations to create and maintain the protocols.

> Now, the IETF may not need to standardize the entire protocol.  So far,
> the drafts that have been submitted are just the basics (stream:stream,
> message, presence, iq, register, auth, roster, I think).  It may be that
> any working group that might be formed might only be interested in the
> basics.

While this may be a viable option, I believe that it would be best to 
keep "the basics" under JSF control, too. This is to allow us to build 
greater and better things on top of Jabber (I refuse to use that "XMPP" 
term) without having our foundation slipping out from underneath us 
while we try to work. This is not a big issue, since I don't see that 
happening, but it might be something to consider anyway.

> The other point I'd like to make is that the IETF is not exclusive.
> There's no fee for participation on the mailing lists, and even though
> there are fees for the IRL meetings, usually those serve as
> ratifications of things that have happened on the mailing lists.  So,
> people who want to make decisions about the protocol will still be able
> to do so, but the words at the top of the document would have to be
> "INTERNET-DRAFT" and not "JEP".

The IETF thankfully does not charge for membership. It doesn't even have 
a membership. This is actually a bad thing because the developers do not 
make the decisions on the future of the IETF. They also don't make the 
decisions on the protocols they are developing, that is by an informal 
consensus of participants, which in the past has shown to take a very 
long time to resolve matters and has caused many disputes and attempted 
forks. Jabber has greatly benefitted from its formal voting and 
membership process. It has kept everyone more or less happy when that 
process is followed, and allowed Jabber to keep moving forward instead 
of having to revisit and refresh past decisions every so often.

> I'd like to point out that in the BOF meeting, I took a lot of heat on
> this issue, since I didn't stand up and say "absolutely, you guys can
> control everything."  What I said was that we have an existing process
> in the JSF, and that we would have to reach consensus within the JSF
> that handing over change control to the IETF is the right thing.

Yes, I've read the Minutes, and thank you alot Joe for taking that heat 
that i read about between the lines :) You did a great service for all 
of us putting yourself on the line like that  :)

> Even though the IETF hasn't said yes to a working group yet, now is the
> time to start the debate about whether having an IETF standard is the
> right thing for us as a community.  I'd say that the answer to this
> should be a clear "yes", since I think it is the best way for the
> protocol to reach as many people as possible as quickly as possible.
> Perhaps there are some really good arguments for not doing it, though.

I know Jabber will come to take over the Internet and take its place as 
the premier universal open protocol for instant communications. It's a 
gut feeling because Jabber has what no other attempt has had: community, 
tolerance, lots of implimentations, and superior technology.

We don't need the IETF to accomplish these things, the only thing the 
IETF would help with is to make this happen a bit faster.

Also, I'm worried about the community of the IETF. Although it is a very 
open organization, I do believe that it has its own feel and informal 
way of doing things that would hinder Jabber's development there. Namely...

With the IETF, I seriously doubt Jabber would come to replace such 
legacy protocols as HTTP and SMTP. The IETF community would more or less 
have a bias towards restricting Jabber to IM and fringe uses, allowing 
HTTP to become the premier protocol for "web services" and the "semantic 
web". I'm not claiming a great conspiracy, just community bias.

Watching a few of the IETF's past Working Groups and hearing about the
strong impression that they create protocols by committee, and few of 
the developers actually try to impliment the protocol they are making. 
This is in sharp contrast with Jabber.

> Here are the questions at hand:
> 1) Should the protocol be standardized by the IETF?
> 2) If so, how much of it?

-- 
     /\  Adam Theo, Age 23, Tallahassee FL USA
    //\\   Email & Jabber: theo at theoretic.com
   //  \\  (Boycotting AOL, therefore no AIM or ICQ)
=//====\\=
//  ||  \\  Theoretic Solutions: http://www.theoretic.com
     ||         "Building Ideas by Bringing them Together"
     ||      Jabber Protocol: http://www.jabber.org
     ||         "The Next Generation Communications Protocol"
     ||  "A Free-Market Socialist Patriotic American Buddhist"




More information about the Members mailing list