[JDEV] Re: jdev digest, Vol 1 #1261 - 11 msgs

[Al Sutton]

I think there are some very important differences between WV and Jabber
that stem from their design criteria, and that any naming similarities
is purely down to common sense being applied in both. The main
differences that come to mind are;

Firstly the use of transactions. Jabber uses the attribute of a tag to
mark a transaction ID which is enough for tracking responses to queries.
WV uses a TransactionDescriptor block that contains not only a
TransactionID sub block, but also a transaction mode (to tell whether
it's a request or reply), and other sub-blocks that describe the
transaction as well as labelling it.

Secondly the use of a SessionType block in WV to describe whether or not
it's a post authentication message. Again this shows a design orientated
around small transmission blocks as opposed to jabbers design
orientation around a single persistent stream.

Thirdly the design is orientated towards more simplistic XML parsers by
not using any attributes (except the namespace declaration). This allows
client developers to implement simplistic XML parsing routines which
don't/can't handle attributes in tags.

Finally, the reason HTTP is used is because it's the commonly
established application layer protocol which is implemented over the WAP
protocol, and thus 'phone producers already have the software technology
to use it.

I think the design goals of jabber (single persistant stream) against WV
(independant message blocks) has resulted in notably different
protocols. 

I'm not saying either is great or bad, I'm just saying both have aimed
at different goals, and that I think we could probably do ourselves a
lot of favours by linking into a protocol that has openly available
specs.

Al.

On Mon, 2002-02-18 at 02:42, Jean-Louis Seguineau wrote:
> One should keep in mind that the WV spec is meant to be used on wireless
> handsets, primarely cellular phones, with SMS and WAP as a bearer.
> And trying to type instant messages that are longer that 100-150 characters
> becomes quickly cumbersome...
> 
> What is interrested however, is that this spec has been largely "inspired"
> by the jabber protocol to say the least. A lot of interresting concepts have
> been added in the field of presence attributes that might find their place
> back into the jabber protocol. The way server-to-server communication is
> handled, and the relay functionality is also worth considering.
> 
> I also tend to aggree with Ian saying that the TCP/IP binding over HTTP is
> not up to the standard, any jabber afficionado will probably aggree. But
> there is nothing in the spec that impose HTTP, it's just an implemetation.
> (BTW why on earth do they all want to use HTTP as a bearer to do real-time
> applications??)
> 
> Jean-Louis
> Antepo, Inc.
> 
> > Message: 8
> > Date: Sun, 17 Feb 2002 15:00:08 -0600 (CST)
> > From: Peter Saint-Andre <stpeter at jabber.org>
> > To: jdev at jabber.org
> > Subject: Re: [JDEV] Specification 1.0 for Wireless Village
> > Reply-To: jdev at jabber.org
> >
> > I like how there's a 3k limit per message.
> >
> > Peter
> >
> > --
> > Peter Saint-Andre
> > email+jabber: stpeter at jabber.org
> > web: http://www.saint-andre.com/
> >
> > On Sun, 17 Feb 2002, Wayne Sheppard wrote:
> >
> > > <snip>
> > >
> > > > At the very least we should keep it on the radar as a possible interop
> > > > target.
> > >
> > > This is one of the primary goals of the spec, assured interop. This is
> > > stated in the WV FAQ, that the hope is that other developers,
> manufacturers,
> > > etc. will 'extend' their offerings to interop with IMPS. It seems to
> have
> > > garnered a lot of support as the founding members cover a lot of ground,
> not
> > > only from a hardware perspective, but from a carrier and technology
> > > perspective as well.
> > >
> > > Cheers,
> > >
> > > Wayne Sheppard
> > > www.corybant.com
> > > www.vastcommunications.com
> > >
> > >
> > > _______________________________________________
> > > jdev mailing list
> > > jdev at jabber.org
> > > http://mailman.jabber.org/listinfo/jdev
> > >
> >
> >
> 
> 
> 
> _______________________________________________
> jdev mailing list
> jdev at jabber.org
> http://mailman.jabber.org/listinfo/jdev