[standards-jig] JNG Ramblings.

Mike Lin mikelin at MIT.EDU
Fri Aug 9 01:36:09 UTC 2002


Hi temas,

I took a look at the OSCAR protocol. Obviously some similar ideas.

I'm basically taking a total crap shoot at trying to find the right
blend of binary framing and XML description so that we get the
efficiency and development ease of the binary side without sacrificing
too much of the basic Jabber philosophy. I'm already cringing a bit at
the envelope/manifesto stuff, which I think is just barely worthwhile.

Right now I think the family & data type are better expressed as XML.
The protocol I laid out has "XML" and "unspecified" binary type codes
because I think it is consistent with our philosophy to have a specific
preference for XML right down to the transport level, while allowing
binary stuff as second-class citizens. Call it racism :-)

So although the data field has 126 other possible values, I don't really
want to use them, which is why I'm putting content-types into the
manifesto.

Anyway, I don't know if any of this is right, but I'm putting my money
where my mouth is and implementing this stuff, so we'll see where it
goes.

-Mike

On Thu, 2002-08-08 at 13:04, Thomas Muldowney wrote:
> On the framing part, have you ever looked at AOL's OSCAR protocol?  It's
> actually rather nifty.  It's composed of a primary wrapper/header they
> call a SNAC.  It defines a family and data type (kind of like namespaces
> in our world).  The SNAC is composed of many TLV's (type, length, value)
> sets.  Which is why I thought of it while looking at your stuff.  I need
> to digest the rest more, but it definately made me think of OSCAR.
> 
> --temas
> 
> 
> On Thu, 2002-08-08 at 10:12, Mike Lin wrote:
> > hi all
> > 
> > i've been coding these past few days on some JNG prototypes. nothing is
> > done yet, but to help get the juices flowing i wrote down some of the
> > things i'm aiming to do.
> > 
> > most of the stuff i've done thusfar relates to the transport layer, and
> > thus slips underneath adam's core tool protocols ideas, although i'm
> > beginning to touch on them near the end.
> > 
> > http://mikelin.mit.edu/xmpp/jng
> > 
> > anyway, i'm just hoping to help get some conversations started. as the
> > <h3> at the top clearly states, this should not be taken seriously. it
> > is not proven to work by any stretch of the imagination. i'm doing some
> > very fluid experimentation in coding it up, and i'll keep things updated
> > with my experiences.
> > 
> > i'm not trying to get everyone to drop everything and switch over to
> > JNG. i just think we need to start thinking about it more formally.
> > 
> > -mike
> > 
> > _______________________________________________
> > Standards-JIG mailing list
> > Standards-JIG at jabber.org
> > http://mailman.jabber.org/listinfo/standards-jig
> 
> 
> _______________________________________________
> Standards-JIG mailing list
> Standards-JIG at jabber.org
> http://mailman.jabber.org/listinfo/standards-jig





More information about the Standards mailing list