Lists Home |
Date Index |
- From: Michael Fitzgerald <email@example.com>
- To: Uche Ogbuji <firstname.lastname@example.org>
- Date: Fri, 22 Dec 2000 11:26:09 -0800
I think you misunderstood what I was asking. I am looking for David's
distinction of "protocol" so I can understand his earlier point, but he is
probably off Christmas shopping (like I should be) and we won't hear back
from him until Jan. 2!
However, I think of SOAP as merely an XML document, not a transport
protocol. Any implementation that processes such a document handles the
transport. Alone it can't do much. SOAP is a driver, not a car. E.g., HTTP
physically moves a copy of content on a server in Kalamazoo to my browser in
Oregon with a GET method. SOAP alone does nothing of the sort.
I also believe SOAP encoding is there as yet another convenience. It's there
if I want it, but I don't have to use it, and the SOAP Body allows me a lot
of flexibility without it.
Best wishes for the holidays,
From: Uche Ogbuji [mailto:email@example.com]
Sent: Friday, December 22, 2000 10:51 AM
To: Michael Fitzgerald
Cc: David Megginson; firstname.lastname@example.org
Subject: Re: RDF vs. SOAP serialization (oh yeah, and XMI and XTM)
> So when you say "protocol," are you referring to the SOAP
> Envelope/Header/Body rather than transport protocols such as HTTP and
When you say "HTTP and SMTP", you don't mean them to be transport
protocols such as TCP, do you? When you say "TCP", you don't mean it to
be a transport protocol such as "IP", do you? When you say "IP"...
Come on. One man's OSI level 3 is another's OSI level 5. Does CORBA'a
CosEvents systems really occupy the same network layer as CORBA's
SOAP is a transport protocol, regardless of the fact that it is layered on
top of other technologies that are also called "transport protocols".
> And when you say "data serialization," are you referring to the encoding
> schema http://schemas.xmlsoap.org/soap/encoding/ alone?
What else? David can speak for himself, but my opinion is that SOAP
should not have defined the Envelope/Header/Body, and then kicked in a
full spec for the format for that body. The latter should have been the
subject of another spec.
Note that RDF made the same error when conflating abstract model and
serialization syntax. I guess it's not unusual for standards-makers to
try to fit the whole world into a kernel.
Uche Ogbuji Principal Consultant
email@example.com +1 303 583 9900 x 101
Fourthought, Inc. http://Fourthought.com
4735 East Walnut St, Ste. C, Boulder, CO 80301-2537, USA
Software-engineering, knowledge-management, XML, CORBA, Linux, Python