RE: XML protocol comparison

Sure, thanks for the clarification.

	Sami


-----Original Message-----
From: Eric Prud'hommeaux [mailto:eric@w3.org]
Sent: Monday, May 08, 2000 2:05 PM
To: Sami Khoury
Cc: XML DistApp ML
Subject: Re: XML protocol comparison


On Mon, May 08, 2000 at 01:36:19PM -0700, Sami Khoury wrote:
> Eric said: "The defined ICE grammer is used only for the transport
protocol,
> and not for defining what is in the payload."
> 
> Right -- that's why I would have classified ICE as generic.  I suspect
that

Every transport mechanism (FTP, SMTP, finger) would fit in that
category of generic.

> by "generic" you meant, "provides a mechanism to represent arbitrary
data",

I'm happy with that definition and incorporated it into the matrix.

> where ICE simply allows an application to transport and negotiate over
> arbitrary data.  However, I don't know of an protocol that provides an
> invertible transform mechanism for arbitrary data (for example, nothing in
> XML-RPC mandates that a GIF be sent as base64 or by reference, if it is to
> occur as a method parameter).

You can take a SOAP message and reconstrcut a graph of inter-related
structures. XML-RPC doesn't do graphs, but you could reconstruct a
simple (address) record that had a name, address, street, etc.

Again, I'm not married to this grouping. I just want folks to
understand why I put thigs where I did (for now).

> -----Original Message-----
> From: Eric Prud'hommeaux [mailto:eric@w3.org]
> Sent: Monday, May 08, 2000 1:29 PM
> To: Sami Khoury
> Cc: XML DistApp ML
> Subject: Re: XML protocol comparison
> 
> 
> On Mon, May 08, 2000 at 12:46:40PM -0700, Sami Khoury wrote:
> > Hm, how are "generic" and "app-specific" being defined?  I ask because
ICE
> > isn't app-specific in any vertical industry sense, nor in any
programmatic
> > sense.
> 
> I suspected my terms weren't the greatest choice, but wanted to get it
> out for discussion quickly. Perhaps I should just sequester those that
> have "non custom serialization". The defined ICE grammer is used only
> for the transport protocol, and not for defining what is in the
> payload.
> 
> WfXML is now in the generic group.
> 
> > -----Original Message-----
> > From: Eric Prud'hommeaux [mailto:eric@w3.org]
> > Sent: Monday, May 08, 2000 10:20 AM
> > To: Bernhard Dorninger
> > Cc: XML DistApp ML
> > Subject: Re: XML protocol comparison
> > 
> > 
> > On Fri, May 05, 2000 at 05:13:06PM +0200, Bernhard Dorninger wrote:
> > > Hi
> > > 
> > > The more I read specs and related material of the "protocols" listed
in
> > > Eric's matrix, th more I feel, that heavyweights like BizTalk, eCo,
> ebXML
> > > should not be mentioned in one go with protocols like XMLRPC, SOAP or
> > WDDX.
> > > The former are far more than just protocols, they provide an
integrative
> > > infrastructure for E-commerce. So IMO BizTalk and Co. should not
> directly
> > be
> > > compared to XMLRPC and Co.,  I think the two "groups of protocols"
have
> > been
> > > designed with completely different intentions.
> > 
> > I was thinking the same thi
> > 
> > Yes, the list seems to benifit from grouping of similar protocols (see
> > http://lists.w3.org/Archives/Public/xml-dist-app/2000Apr/0058.html). I
> > meant to propse furthur grouping, but apparently failed.
> > 
> > I just made a quick pass at this (see
> > http://slow1.w3.org/2000/03/29-XML-protocol-matrix). Summary:
> > 
> > generic:
> >   XML-RPC
> >   SOAP
> >   WDDX
> >   XMI
> >   jabber
> >   ebXML
> >   BizTalk
> >   BXXP
> >   LOTP
> > 
> > app-specific:
> >   ICE
> >   IOPT
> >   WfXML
> >   eCo
> >   XMOP
> >   
> > non-XML:
> >   TIP
> >   XDR
> >   HTTP-NG
> >   template
> >   
> > I haven't read all of these specs so some of these may be in the wrong
> > place. Pleast post corrections to the list.
> > 
> > -- 
> > -eric
> > 
> > (eric@w3.org)
> 
> -- 
> -eric
> 
> (eric@w3.org)

-- 
-eric

(eric@w3.org)

Received on Monday, 8 May 2000 17:15:09 UTC